Author: Site Editor Publish Time: 2026-09-16 Origin: Site
Magnetorheological finishing makes local material removal a measurable, model-based and iteratively verifiable process in precision optics. “Deterministic” means that measured surface error can guide dwell-time calculation when the machine state, workpiece material, process conditions and removal function have been established and validated. It does not mean that the final result is guaranteed without measurement. [1–3]
A correction loop typically includes surface measurement, data conditioning and error definition, removal-function calibration or verification, dwell-time calculation, CNC execution, post-process measurement and any further correction. This loop can alternate machining with independent metrology. In-machine measurement or automatic feedback during processing is a separate equipment-configuration question.
Before finishing begins, the process must define what material needs to be removed. For optical surfaces, the target may be derived from interferometric or other suitable metrology. The valid aperture, reference surface, detrending method and spatial-frequency content to be corrected should be defined. Otherwise, different data-processing rules can produce different correction targets for the same physical surface.
A signed surface-error map must be converted into a nonnegative removal target using the reference surface, any permissible overall offset and the available stock. Removal cannot add material to low areas; extra removal must respect thickness, geometry and other drawing constraints. Establish whether the input represents surface height, optical path difference or wavefront data, and use a conversion consistent with the measurement geometry and instrument definition. [3–5]
ISO 10110-5:2026 addresses drawing indications for optical surface-form tolerances and prefers nanometres for surface-form deviations; fringe spacings remain permitted with an explicitly stated base wavelength. ISO 14999-4:2026 addresses interpretation and evaluation of surface-form and wavefront-deformation tolerance data. These are specification and evaluation frameworks, not equipment-performance certifications. The applicable editions and acceptance method should be identified for each project. [4,5]
A removal function describes the spatial distribution of material removed per unit time under defined process conditions. It is not a universal constant. Its shape and magnitude can depend on magnetic-field conditions, polishing-fluid state, workpiece material and geometry, tool size, relative motion and other process variables.
A removal function can be established using a suitable specimen or a workpiece on which calibration is permitted, or predicted with a validated model. Calibration or validation should cover the relevant material, geometry and process state. A removal rate from another machine or material should not be transferred without validation. [1]
Calibration records should identify the fluid system, batch and actual condition in use. After replacement, replenishment, conditioning or a change in operating state, follow a validated procedure to confirm that the removal function remains applicable and recalibrate where needed. Rate values also need defined units and meanings: peak depth-removal rate, volumetric removal rate and whole-part cycle time are not interchangeable. A value in λ/min requires the reference wavelength and the surface-height or wavefront basis to be stated.
In a discrete model that approximates removal as a linear superposition of local contributions, predicted removal can be written as:
Here, h with the subscript pred is the predicted removal-depth distribution. Each column of A represents the removal per unit time at the corresponding tool position under specified process conditions. The vector t contains the equivalent dwell times, each of which must be nonnegative. If A is expressed in nm/s and t in seconds, predicted removal is in nm. The solver seeks to approximate a required nonnegative removal distribution; a signed surface-error map is not itself that distribution.
Neighboring tool positions have overlapping influence areas, so dwell time is normally obtained through inversion or optimization with appropriate regularization, boundary, path, speed, acceleration and stock-removal constraints. A convolution is a further simplification that requires an approximately shift-invariant removal function and sufficient stability during processing. A numerically feasible dwell map must also be executable by the motion system. [2,3]
Pre-process measurement establishes the correction target; post-process measurement checks the actual result. Differences between prediction and measurement can arise from removal-function drift, fluid state, workholding and registration errors, measurement noise, edge effects, machine motion errors or material response. Further correction should take account of the permitted residual, available stock, measurement repeatability and uncertainty. [6]
· Define the valid aperture and reference surface explicitly.
· Use comparable registration, sampling, detrending and filtering rules, and record the evaluated spatial bandwidth.
· Evaluate form using peak-to-valley deviation (PV), form RMS, residual distribution and drawing requirements. Roughness parameters such as Ra or Sq are different quantities and do not replace form acceptance.
· Do not confuse filtering, edge treatment or metrology artefacts with physical machining results.
1. Measure the current surface and record the metrology conditions.
2. Define the evaluation aperture, nonnegative removal target and available stock.
3. Confirm the removal function for the relevant machine, material, geometry and fluid state.
4. Solve for dwell time and tool path with nonnegative dwell, boundary and motion constraints.
5. Execute the correction with controlled process conditions.
6. Re-measure the surface using a comparable metrology procedure.
7. Compare the measured residual with acceptance requirements, considering uncertainty and remaining stock.
Deterministic methods improve quantitative correction based on measurement, but results remain dependent on metrology, modelling, the equipment, the material and process stability. PV, RMS, roughness or processing time from one specimen is not a universal guarantee for other workpieces. A high removal rate also does not automatically imply better surface texture or a shorter total production cycle.
For an actual project, equipment selection should consider workpiece material, clear aperture and curvature, freeform characteristics, required surface form and texture, available metrology, process throughput and the customer's acceptance method. Maximum workpiece size or a single headline accuracy figure is rarely sufficient to determine process suitability.
APEX supports magnetorheological finishing equipment selection and project coordination. Configuration, application scope and target performance should be established through the relevant model documentation, workpiece-process evaluation where needed, and mutually agreed written acceptance criteria. Share your workpiece material, drawing, starting surface condition and target requirements with APEX to discuss an equipment project.
A reliable correction loop connects measurement, an executable removal model and post-process verification. The applicability of the removal function, accurate dwell execution and comparable metrology determine whether the correction result can be interpreted with confidence.
[1] Dai, Y., Song, C., Peng, X., and Shi, F. (2010). Calibration and prediction of removal function in magnetorheological finishing. Applied Optics, 49(3), 298–306. Source
[2] Song, C., Dai, Y., and Peng, X. (2010). Model and algorithm based on accurate realization of dwell time in magnetorheological finishing. Applied Optics, 49(19), 3676–3683. Source
[3] Li, L., et al. (2017). Positive dwell time algorithm with minimum equal extra material removal in deterministic optical surfacing technology. Applied Optics, 56(32), 9098–9104. Source
[4] ISO 10110-5:2026. Optics and photonics — Preparation of drawings for optical elements and systems — Part 5: Surface form tolerances. Source
[5] ISO 14999-4:2026. Optics and photonics — Measurement of optical elements and optical systems — Part 4: Interpretation and evaluation of surface form and wavefront deformation tolerances specified in ISO 10110. Source
[6] JCGM 100:2008. Evaluation of measurement data — Guide to the expression of uncertainty in measurement. Source
This article provides general technical information and does not replace workpiece-specific process documents, inspection reports or safe operating procedures. Equipment configuration, application scope and acceptance criteria are governed by the relevant technical documents and written agreement. This article creates no additional performance warranty and does not alter obligations imposed by applicable law.
