1. Separate contact-surface and assembly functions

Distinguish the skin-contact region, visible exterior and assembly region seating against an internal component or housing. These regions belong to one metal head but require different geometry, finishes and inspection references.

The broad central contour and small perimeter radius are distinct features. The central contour influences contact direction and area; the edge defines the local transition into the sidewall. An instruction to make the whole head rounded does not adequately define both.

Choosing flat, gently convex or segmented contact surfaces depends on the application area, contact direction, movement and operating mechanism. Greater convexity or a larger radius does not always improve contact. Assess the broad face and perimeter’s relative positions in the intended operating orientation.

Geometry deviations that change contact

Components with the same overall outline can first touch at different locations if their central height or peripheral bow differs. Local height differences also affect load distribution. Actual contact area varies with skin deformation, applied force, inclination and application area; the projected area of the metal head must not be treated as its actual contact area.

Review electrical, thermal and mechanical operating conditions alongside geometry. Component-shape inspection alone does not demonstrate complete-device output uniformity or performance on skin. Define the required geometry on the component drawing and actual contact and operating conditions in complete-device evaluation.

Continuity between surfaces

Check sections for unintended ridges or flat regions between the central face and edge. Examine CAD tangent continuity, parting-line steps after molding and local flattening after polishing. Reflected lines help locate irregularities but do not replace dimensional assessment.

Central height versus circumferential form

Discs with identical diameter and central height may have a raised edge or dish-shaped perimeter. Compare radial sections with circumferential sections at fixed radii to distinguish tilt, central convexity and local edge recession. One central height and an average deviation cannot establish the entire contact surface.

If the assembly seat is datum A, assess the functional profile from that datum. Freely best-fitting the measured point cloud to CAD can absorb tilt or height error. Distinguish shape-only alignment from assembly-position verification, and implement the drawing’s datums and constraints in the inspection program.

2. Material, feedstock and sintered microstructure establish the mirror-finish foundation

Higher gloss requires control of polishing and the underlying metal microstructure. Contact face, thin sidewall and internal fasteners can have different requirements. Select material from the contact environment, fastening loads, final finish and process sequence rather than appearance alone.

Choosing between 316L and 17-4PH

316L is a MIM candidate where corrosion resistance, toughness and polishability matter together. 17-4PH is a candidate for higher-strength structures such as fasteners or a thin neck; specify its properties with the precipitation-hardening condition. Being stainless steels does not justify identical processing or acceptance criteria.

A bright finish does not establish corrosion resistance or contact suitability. High hardness alone likewise does not establish suitability. Compare whether repeated fastening loads or broad mirror-surface appearance governs the design, and whether the finished surface meets the intended environmental requirements.

Feedstock uniformity beyond powder fineness

Finer powder can increase the sintering driving force, but size alone does not determine quality. Particle-size distribution and shape influence packing and binder demand; powder-to-binder ratio affects viscosity and filling. Insufficient binder can impair flow; excess binder can promote separation during molding and impair shape retention after debinding.

Where flow changes from a broad contact plate to a narrow connection, confirm uniform composition throughout the filled cavity. Local powder-content differences can create nonuniform sintering shrinkage. Correcting only the overall mold scale can leave contact-surface bow or circumferential height differences.

When introducing a new feedstock, review the supplier’s shrinkage factor and flow data, then connect actual mold-filling behavior, green-part mass and post-sintering dimensions at defined locations. Constant total mass does not automatically demonstrate uniform internal composition. If a deviation recurs in a particular region, examine its cross-section and, where needed, local density and microstructure.

Debinding controls shape and final composition

Debinding must allow binder removal while preserving the powder skeleton. With thick bosses joined to a thin plate, the removal path through the thickest region matters more than average thickness. Subsequent polishing cannot repair debinding cracks or internal damage.

Residual constituents also matter. In MIM 17-4PH in particular, residual carbon after debinding and sintering can affect phase fractions and the response to heat treatment. Debinding mass loss can be used as a process-tracking indicator, but it does not by itself establish final carbon content or material performance. Connect material certificates, any necessary composition analysis, sintered microstructure and heat-treatment records.

Average density versus surface pores

Increasing sintering temperature and time can promote both densification and grain growth. Higher temperature therefore does not ensure simultaneous improvements in strength, dimensions and mirror finish. Define the process window from densification, microstructural change and support-related distortion together.

Similar bulk densities can conceal different pore sizes, shapes, distributions and surface connectivity. Open pores affect electrolyte exposure and local corrosion, so average density and low roughness alone cannot qualify a mirror surface. Inspect final-surface pinholes and examine sections where needed.

3. Design mold filling and edge finishing together

Deburring controls unintended projections and sharp remnants; a target radius defines designed geometry. A “remove burrs” note alone does not specify a consistent radius or its transition into the central contour.

Identify edges touching skin or potentially contacting it during movement, and define their sections. Specify radius ranges and inspection sections where needed. Treat burrs, parting-line steps and gate-removal marks separately. Do not assign a universal safe radius or roughness to all beauty devices.

Broad-plate filling and flow-front junctions

If a broad contact plate and narrow connection form one MIM part, their flow sections and filling conditions differ. Reducing gate area at constant volumetric flow changes local velocity, shear and pressure loss. Confirm that a smaller visible gate mark does not compromise filling uniformity. Flow fronts passing around bosses or holes may leave weld marks. Use flow analysis and staged short shots to inspect last-to-fill regions, junctions and air-escape paths.

For incomplete filling, review gate location and area, mold temperature, filling speed and feedstock flow range rather than pressure alone. Apparently complete filling does not ensure uniform powder distribution or residual stress. Molding stresses can relax during debinding and cause distortion. Compare geometry immediately after injection with orientation and profile after debinding to separate molding history from sintering deformation. Link final distortion locations to molding flow paths.

Gate and parting-line placement

Locate witness marks with removal and finishing access in mind. Removing marks across the contact face or curved transition can alter surrounding geometry. Prefer locations outside the contact face where possible while meeting flow, filling and demolding requirements.

A rear witness mark can still interfere with an assembly datum or internal support. Distinguish contact, datum and fastening faces before selecting mold orientation and gate-removal access. Hiding a mark is different from eliminating its functional interference.

4. Control shrinkage magnitude separately from distortion direction

Depending on design, the head’s back may contain fastening bosses, component seats and locating features. Thick regions can fill, debind and sinter differently from the surrounding thin contact face. Stable exterior curvature therefore starts with internal sections.

Thickness transitions and material concentration

Maintain reasonably uniform thickness and reduce abrupt transitions. Consider coring bosses while preserving required fastening length and surrounding wall. Root fillets may be necessary, but excessive reinforcement should not create a thick mass.

Supporting ribs require similar care: more ribs alone do not ensure stability. Rib-intersection thickness, filling direction, sintering support and machining-fixture locations must work together.

Shrinkage compensation versus sintering support

For the actual cavity, include molding dimensional changes, cavity variation and machining allowance. Supplier shrinkage data are a starting point; verify directional and local shrinkage on actual-geometry process samples. Check whether the factor already covers total cavity-to-sintered shrinkage, and do not count molding shrinkage twice.

Overall enlargement establishes target size but cannot correct local bow, tilt or asymmetric distortion. Identify deformation positions and directions in repeated samples, then adjust geometry, process conditions and sintering support together.

Setter positions and contact conditions can affect shape. Check for final-surface contact marks, constrained shrinkage and unsupported spans. The underlying relationships also appear in MIM sintering shrinkage and mold compensation design.

Peripheral sag versus support marks

Center-only support can leave problematic unsupported outer spans. Perimeter-to-setter friction can obstruct free shrinkage by restricting movement. A larger support area alone is insufficient: review contact locations, relative movement, friction and shrinkage direction.

If parts from one cavity distort differently with loading orientation or setter position, isolate support conditions before correcting cavity dimensions. Distinguish cavity-related from batch-related deviations to identify the principal source among tooling, molding and thermal processing.

5. Selective machining connects assembly datums and contact-plate stiffness

After MIM forms the exterior and internal geometry, selected functional faces can be machined or ground. Candidates include height-setting seats, boss ends and mating-part locating holes. Identify assembly-critical regions before assigning identical machining tolerances everywhere.

Connect inspection and machining datums

If boss machining locates from the outer profile but assembly locates from another seat, additional variation enters the contact-height relationship. Connect machining, inspection and assembly datums on the drawing.

Thin-plate stiffness and unclamped shape

For a flat, uniform, isotropic thin plate within the linear-elastic range under small-deflection conditions, bending stiffness is expressed as D = Et³ / [12(1 − ν²)]. E is Young’s modulus, t is thickness and ν is Poisson’s ratio. For the same material, stiffness decreases with the cube of thickness. Review polishing allowance and final remaining thickness together with deformation under a given clamping load.

Actual heads contain curvature, bosses, ribs and steps, so this equation cannot establish finished-part deformation. Include support positions and clamping loads in analysis and sample comparison. Machining a distorted part while forcing it flat can change its released shape. Distinguish machining-datum repeatability, unclamped free state and final fastening state.

Repeatable seating and functional-feature sequence

A controlled outside diameter is insufficient if another face determines assembly height. Use the seat for height and tilt, locating geometry for the center, and a rotational constraint where needed. Match fixture contacts to drawing datums rather than arbitrarily pressing different protrusions on different parts.

Where boss ends, locating holes and contact-face positions interact, share datum setups wherever practical. Each setup adds relocation variation. Managing datum-transfer frequency and repeatability can be more effective than tightening machining tolerances alone.

Allowance for later processes

Machined dimensions may differ from delivery dimensions: polishing removes material and surface treatment can change size. Identify finishing regions and datums to preserve, separating exterior-form allowance from assembly-dimensional allowance.

Plan selective machining and fixtures with datum setting for precision CNC machining. Define how much sintered variation machining must correct to reduce unnecessary full-surface machining and polishing removal.

6. Define the mirror-polishing stages

Separate form, waviness and roughness

On a circular mirror-finished plate, brightness and surface accuracy differ. Fine-mark removal can increase gloss while leaving overall dish-shaped bow or intermediate-scale undulations. Assess design-profile deviation separately from waviness and roughness within defined wavelength bands. Parts meeting identical Ra requirements may still differ in reflected-line continuity and perimeter contact geometry.

“Glossy” alone cannot specify a mirror finish. Match reference-sample lighting and viewing direction, and define acceptable scratches, haze and point defects. Separately control plate profile, edge sections and final height to detect functional geometry lost while adjusting gloss.

Shape correction and final gloss

Prolonged soft buffing of deep machining marks can remove peaks while leaving grooves. First remove previous-process marks by grinding or form-supporting polishing, then progress to finer abrasives. Remove residual abrasive between stages to prevent coarse particles from scratching the fine-polished surface.

Mechanical polishing removes material through abrasive-tool-surface contact. Flexible buffing wheels or cloth can wrap around the perimeter; resulting edge roll-off may escape a central roughness check. Inspect center-to-edge sections on circular plates and inspect rib crests separately from grooves.

Identical machine force and rotational-speed settings do not ensure identical local removal. Curvature and tool deformation change contact area and pressure distribution, while overlapping tool paths and edge crossings change local contact duration. This approximate relationship helps explain differences between conditions; predicting removal on an actual head requires trial processing to establish the coefficient and applicable range. It must not be applied directly to electropolishing removal calculations.

Electropolishing versus passivation

Electropolishing removes material by electrochemical dissolution with the part as anode. It may reach surfaces inaccessible to mechanical tools, but results depend on current distribution, electrode arrangement and initial condition. Height and edges can change as microtexture improves. Review allowance and geometry rather than automatically substituting it for mechanical polishing.

Chemical passivation of stainless steel addresses foreign iron contamination and the development of an appropriate passive surface condition. Its purpose differs from a mirror-finishing process that levels deep machining scratches or corrects bow. Visible gloss after finishing does not establish the absence of residual abrasive, oil or foreign metal contamination. Plan cleaning, any necessary surface treatment and verification of treatment results separately.

Exposed pores versus overpolishing

Removal can expose near-surface pores in MIM material; inclusion pull-out and local excessive removal can also create point defects. Do not classify every dot as a sintering pore. Distinguish pores, pull-outs and residue by magnification and, where needed, sections. Bulk density alone does not describe local contact-surface porosity.

If continued polishing produces irregular roughening resembling orange peel, simply increasing pressure and time is inappropriate. Review the material microstructure and hardness, tool compliance and removal conditions at each stage. Separating overpolishing from exposure of pre-existing pores is necessary to distinguish defects suitable for repolishing from those requiring correction at the material or sintering stage.

Different paths for circular faces and repeated ribs

Control a disc’s broad face together with its circumferential transition. Ribbed curved heads have different crest and groove contact conditions. Excessive crest dwell reduces rib height while grooves may retain earlier marks. Adapt tool size, stiffness, access and path to geometry even when the gloss target is identical.

Visible surfaces and assembly datums need not receive identical polishing. Define exposure and fixtures to achieve contact-face and edge finish without unnecessary seat, thread or locating-feature removal. Final polishing should finish previously established geometry, not conceal initial shape defects.

See also MIM mirror-finish quality, surface pores and polishing. Compare pre-/post-treatment profiles and removal to establish process suitability rather than specifying only a treatment name.

7. Include polishing removal in the final assembly-step calculation

Gap width and surface height difference at the metal-to-housing boundary are separate quantities. Determine flushness, intentional protrusion or local tilt from assembled relative heights. Define the intended relationship from actual function first.

Height calculation including polishing removal

From pre-polishing dimensions, a simple axial model is Δh = H₀ − a + b − P. H₀ is pre-polishing metal-seat-to-contact height; a is axial contact-face removal; b is common-datum-to-metal-seat assembled height; P is height from that datum to the adjacent housing. The model assumes the seat is preserved during polishing, with no tilt, fastening deformation or added surface layer.

If finished height Hf is already measured after polishing, use Δh = Hf + b − P. Subtracting a again or re-adding its variation counts the same process twice. Distinguish finished drawing dimensions from intermediate dimensions before selecting the equation.

Why one representative height is insufficient

Correct central height does not establish peripheral or edge height. On circular heads, define several central sections and circumferential comparison points; on elongated heads, inspect lengthwise, transverse and edge-transition sections. Averaging does not eliminate tilt or bow.

Screws or snaps can alter seating and deformation. Accepting boss ends alone can miss variation between fastening locations. Distinguish individual-part inspection from checks reproducing final fastening conditions.

Statistical tolerance conditions

Root-sum-square combination without established distributions and independent variation can underestimate adverse combinations. With limited data, examine limit combinations first. Assess distributions, correlations and process-center shifts before adopting a statistical model.

8. Verify form, texture and local defects separately

Contour assessment compares intended geometry with actual shape. Under specified measurement and filtering conditions, texture separates into longer-spaced waviness and finer roughness. Low Ra does not establish correct overall curvature or edge radius.

Form removal versus form information to retain

Nominal-form removal and filters affect roughness on a curve. When inspecting contact-surface bow, do not remove that bow as background form. Choose alignment and form-removal methods for the quantity being evaluated.

Align contour inspection to drawing datums. A freely optimized best fit can conceal assembly-relevant positional error. Distinguish datum alignment from shape-only best fitting.

Optical measurement coverage

A mirror surface is not inherently impossible to measure optically. Under suitable conditions, interferometry can measure smooth, highly reflective surfaces. However, as the slope increases at the circumferential transition, reflected light can leave the objective’s collection range and measurement points can be lost. A higher-numerical-aperture objective can increase slope acceptance, but may constrain field of view and working distance.

Central-surface data cannot qualify an unmeasured edge. Identify valid coverage and missing points, changing part orientation or combining measurement principles where necessary. Distinguish interpolated regions from acquired data to support appearance with quantitative form verification.

Ra versus local defects

Ra averages absolute roughness-profile height deviations over the evaluation length. A deep narrow scratch and distributed fine texture can yield similar averages. Record scratch, pinhole and edge-damage locations and sizes separately; select additional profile or areal parameters by function.

Measurements across and parallel to the lay can differ. For before/after comparisons, match location, direction, form removal, filters, evaluation length and instrument settings. Profile and areal measurements are complementary; Ra and Sa are not interchangeable.

One roughness scan does not automatically verify an edge radius. Consider accessible sections, stylus geometry and optical slope acceptance. Combine profile measurement and magnification to distinguish burrs, roll-off and radius errors.

Measurement repeatability versus process capability

Repeatedly operating an instrument at the same location does not sufficiently include the effects of refixturing and re-establishing datums. Repeat measurements with removal and reseating, as in the actual inspection process, and include operator and measurement-day conditions to distinguish instrument variation, datum repeatability and long-term changes. If measurement variation is comparable to the allowed deviation, it becomes difficult to separate differences between components from differences introduced by inspection.

First examine production data by cavity, sintering batch, fixture and polishing condition. Mixing groups with different centers can create an overall distribution with several peaks. Interpret Cpk = min[(USL − μ)/(3σ), (μ − LSL)/(3σ)] only when the process is stable and the distribution assumptions are appropriate. USL and LSL are specification limits; μ and σ are the mean and standard deviation of that process. Sample estimates contain uncertainty, so acceptance of a few initial parts or a single high index must not be converted into a long-term quality guarantee.

9. Trace defect origins before correcting production

Similar final defects can originate in different processes. A mirror-surface dot is not always a sintering defect; circumferential height variation is not always a mold-dimension issue. Connect measurements after sintering, functional machining, intermediate polishing, final finishing and assembly to determine suitable rework.

Initial-production process references

Confirm repeatability across produced parts rather than choosing the best-looking sample. Trace cavity, feedstock batch, sintering loading, machining fixture and polishing path to connect increased variation with its location and process stage.

Retain intermediate profile and seating-height measurements to identify process changes. Compare appearance using agreed lighting, viewing directions and defect criteria. Assess different measured quantities separately: high gloss does not prove accurate form, and low Ra does not prove absence of pinholes.

Connect drawings, finish and quantity

A manufacturing review requires external 3D geometry, a datum-defined 2D drawing, contact zones, final finish, internal fastening geometry and the height relationship to the adjacent housing. Defining the material and heat-treatment condition, selectively machined regions, seats to be preserved, appearance-comparison conditions and production quantity together distinguishes what must be achieved in the mold from what must be completed in later processes.

Tighten tolerances according to function. Connect machining and measurement datums for assembly-height faces, and sintered quality with staged removal for broad mirror surfaces. Applying identical mirror finishes and tolerances to less demanding hidden faces adds time and inspection effort without necessarily improving function.

MIM beauty-device head quality connects material and sintered microstructure with geometry-preserving polishing and final assembly datums. Define gloss, profile, edges and height separately, then control changes between processes for repeatable production.