Die casting supports tight tolerances for motion-critical parts through stable tooling, controlled thermal and shot conditions, feature placement, selective CNC machining, a coherent datum scheme, capable measurement, and cavity-specific production studies. It does not ensure robot accuracy by itself. Bearing alignment, gear mesh, encoder position, assembly preload, structural compliance, temperature, calibration, and control behavior must also be validated on the assembled axis.
Identify the output requirement: end-effector pose, axis repeatability, backlash, straightness, runout, sensor alignment, gear contact, rail motion, or process path. Allocate acceptable error across bearings, shafts, gears, joints, casting, machined features, fasteners, fixtures, assembly, temperature, controls, and calibration. Do not assign every drawing feature the same tight tolerance.
Define loads and orientations. A housing may align at inspection and deflect under payload or motor torque. Gravity, acceleration, bearing preload, gearbox reaction, cable force, and temperature can move datums. Functional limits should protect performance over the stated duty rather than only a free-state layout.
Features formed in one die member can have a different capability from features split by the parting line, slides, or core actions. Draft, ejection, gate and overflow removal, thermal gradients, thin walls, heavy bosses, and tool wear affect geometry. Locate precision stock and datums where filling, cooling, ejection, trimming, and measurement remain controlled.
Tool design should identify bearing-forming inserts, slides, cooling circuits, ejectors, replaceable wear zones, and maintenance access. Tool steel name does not guarantee dimensional stability. Actual thermal state, geometry, machine, process window, maintenance, and repair determine the result.
Control material identity, melt condition, transfer, die thermal balance, shot profile, vacuum where used, spray, cooling, cycle interruptions, ejection, trim, and handling. These inputs affect fill, shrinkage, residual stress, flash, distortion, and machining stock. Monitor parameters that are causally connected to the characteristic and define reaction limits.
Warm-up and restart need explicit rules. Selected samples after stabilization can hide early drift. Tool maintenance or a repaired insert may shift a datum without changing the nominal process. Identify output by cavity, tool revision, insert, machine, and relevant process state so containment is possible.
Post-machining commonly establishes bearing bores, motor pilots, gearbox centers, encoder seats, rail mounts, seal lands, threads, and datum planes. Plan enough stock without cutting into high-risk internal zones. Fixture location and clamping must represent stable casting features and avoid flexing a thin housing into conformance.
Account for residual stress, porosity breakout, tool wear, burrs, coolant, chips, washing, coating, inserts, and final assembly. Multiple setups introduce datum transfers. A tight bore diameter is insufficient if its axis, shoulder, mating center, and loaded housing relation are not controlled.
Motion interface | Primary control | Final evidence |
|---|---|---|
Paired bearing bores | Common machining datum and stable fixture | Bore size/alignment, preload and loaded rotation |
Motor-to-gearbox center | Related machined pilots and shoulders | Gear contact, torque, noise and backlash |
Encoder or vision mount | Controlled datum plus assembly method | Pose, calibration residual and thermal drift |
Rail or guide surface | Machining, support and fastener sequence | Straightness, friction and path behavior under load |
Housing closure and seal | Final coating state, flange and torque | Fit, gap, ingress/leak and service test |
Define alignment, datum targets, free or restrained state, support, temperature, filters, evaluation zone, coating state, and decision rule. CMM, optical systems, scans, gauges, roundness instruments, surface measurement, and functional fixtures answer different questions. X-ray can inspect some internal conditions but does not measure every functional tolerance or guarantee performance.
Use measurement-system studies appropriate to the risk and data type. Correlate supplier and customer methods before approval. Flexible housings can change under fixture force; rough cast surfaces can change probing results; software alignment can hide a localized shift. Resolve those differences in the specification.
Study stable production-intent output using the intended material, tool, cavities, machine, trim, machining, finish, gauges, and rate. Keep cavities separate until their location and variation are understood. Review distribution, drift, maintenance, warm-up, measurement uncertainty, and customer calculation rules before interpreting capability indices.
Set ongoing frequency and reaction from risk. A launch study does not cover future tool wear or repair. Reverify after changes to material, machine, cavity insert, cooling, gate, process window, trim, fixture, toolpath, coating, gauge, software, site, or sub-tier.
Inspect dimensions, then assemble with production bearings, gears, shafts, fasteners, lubricant, motors, encoders, seals, and cables. Measure friction, torque, backlash, runout, alignment, calibration, repeatability, path, vibration, noise, temperature, and drift under relevant load and duty. Calibration can correct repeatable geometric error but cannot stabilize a loose joint or thermally moving housing.
Send the error budget, datums, loads, temperature, critical characteristics, final-state definition, mating parts, machining, coating, measurement, capability, assembly, tests, traceability, and changes. Ask for DFM exceptions, tool relationships, cavity plan, fixtures, methods, and evidence. Tight tolerance is a controlled chain from drawing to axis behavior, not an isolated number on a casting quotation.