Gearbox housings arrive in the assembly area with residues that are easy to miss and expensive to ignore: machining oil, way-lubricant film, metal swarf, grinding paste, and the sealant or thread-locker applied upstream. Once a bearing, shaft, or gear is pressed into the housing, cleaning defects become latent failures. Acceptance criteria are how a facility turns "looks clean" into a repeatable pass/fail decision that purchasing, process engineering, and the supplier all interpret the same way.
Start with the defects that actually cause field problems
Before choosing any measurement method, list the failure modes specific to your gearbox design. In enclosed gear drives, the usual offenders are:
- Retained metal fines and swarf, which circulate in the oil film and accelerate tooth wear and bearing brinelling
- Cutting fluid and way-oil films that dilute the gear oil, destabilize additives, and interfere with seal adhesion during assembly
- Grinding and lapping compounds trapped in cored cavities, bolt holes, and rib undersides
- Paint, primer overspray, or rust-preventive oil on mating faces where a gasket or seal must seat
- Hand, glove, and skin residues left after manual deburring and handling
Each defect type points to a measurable criterion. Write criteria against the defects you named, not against a generic cleanliness phrase borrowed from another industry.
Define the criteria across four dimensions
A usable acceptance spec for a housing states four things. Missing any one of them leaves inspectors to improvise.
- Contaminant limits. What is allowed: total particulate mass per housing, largest permitted particle size, and whether residual oil film is acceptable or must fall below a stated level. State the contaminant by type, because "clean" after a machining operation means something different from "clean" after painting.
- Zones and priority surfaces. A housing is not uniformly critical. Bearing bores, seal grooves, gasket faces, and oil-gallery bores deserve tighter limits than an external non-machined face. Divide the part into zones in the drawing or work instruction, and attach a limit to each zone.
- Test method and sampling. Name the verification method per criterion — gravimetric wipe or flush for particulate mass on accessible machined surfaces, solvent wipe and gravimetric determination for oil film, particle-count analysis on a flush of the internal cavities, and visual inspection against a reference for areas where flushing is not practical. Specify sample size and frequency, and specify how many housings per lot or per shift are verified.
- Disposition rules. State what happens on failure: rewash and retest, quarantine the lot, or trigger a process review after a set number of consecutive failures. Without this, criteria exist but decisions vary by inspector and by shift.
For an OEM purchase, hand the supplier the same four-part spec rather than a request for "clean parts". A supplier cannot bid reliably against an acceptance rule they cannot measure.
Match the cleaning process to the criteria, not the other way around
Once limits exist, process selection follows logically. Heavy viscous residues and grinding compound need heat and pressure; light oil film may need only a hot rinse and a controlled dry-down. If your acceptance criteria include residual oil on a seal seat, a rinse that leaves oil behind will never pass, regardless of how long the cycle runs.
High-temperature steam is a practical option where housings are heavy, partly machined, and difficult to immerse, and where the process must run at the line or in a maintenance bay instead of in a wash tank. Steam delivers heat for grease and viscosity reduction with far lower water volume than a flood wash, and dry-steam mode limits the standing liquid that carries loosened debris back into blind cavities. For that duty, review the industrial steam cleaning equipment in PowerDe's cleaning series; the fixed PowerDe HW02 steam cleaner is a 220 V unit rated at 6 kW heating power, 160 °C steam temperature, and 0.8 MPa vapor pressure, with an 80 L SUS304 tank and selectable dry or wet steam. Fixed steam cleaning suits an in-plant housing-cleaning station; mobile and hand-pushed units fit low-volume, multi-site, or maintenance work. If your housings have geometry that no standard machine clears — deep bores, internal galleries, or fixtures that need indexing — customized steam cleaning equipment from an intelligent cleaning equipment manufacturer is the route to evaluate, and it should be specified against your acceptance limits, not a catalog brochure.
Prove the criteria before you enforce them
A criterion that no cleaning process can meet will be quietly ignored within a month. Run a qualification pass on your own parts:
- Take parts as they arrive from machining, and establish the baseline contamination level per zone.
- Clean candidate housings with the intended method, then test with the stated method and measure the result against your draft limits.
- Repeat across the machine, tool, and shift variation that actually exists in production. A criterion that passes on a sample part but fails across three shifts is not a criterion, it is an aspiration.
- Verify the measurement itself: same wipe material, same flush volume, same filter, same balance, same operator technique. Two labs reporting different numbers on the same part is a method problem, not a cleanliness problem.
- Then set limits at a level your qualified process reliably achieves, with margin for drift.
Document the qualification as a short report tied to the part number. It is the reference used when a supplier disputes a rejection, and when a new inspector joins the line.
Common mistakes in housing acceptance specs
Several patterns recur, and each of them creates a spec that looks rigorous but does not control quality:
- Visual-only acceptance. Oil film and fine particulate in a blind cavity are invisible on a shop floor under ordinary lighting, and the defect surfaces later as premature wear.
- One blanket limit for the whole housing. The result is over-cleaning of non-critical surfaces and under-cleaning of bearing bores.
- Criteria without a stated test method. Two inspectors, two results, and no defensible rejection.
- Cleaning before deburring. Swarf removed from a machined face is replaced by swarf generated at the deburring station, so the housing fails again at final inspection.
- Ignoring the dry-down. A housing that passes cleanliness but carries trapped water into assembly invites flash rust before the gear oil is charged.
- No drift plan. Filters, nozzles, and wipe materials degrade. Without periodic re-verification, an acceptance spec decays while the paperwork stays clean.
What to ask a cleaning-equipment supplier for
When evaluating equipment against your acceptance criteria, request the following in writing:
- Steam temperature and pressure, heating power, tank volume and material, and voltage, stated as published parameters — the figures for the HW02 above are a benchmark for what should be disclosed.
- Whether dry-steam and wet-steam modes are selectable, since the two behave very differently in blind cavities.
- Fixture and nozzle arrangements for your housing geometry, and whether the supplier will run a trial on your parts before you commit.
- Protection features on the machine and the service terms for wear parts.
- A written capability statement against your numeric limits, so the supplier is accountable for the criterion rather than the machine.
Bottom line
Acceptance criteria for gearbox housings work when they name the contaminant, the zone, the measurement method, and the disposition on failure — and when a qualified cleaning process can meet those numbers day after day. Set the limits from your own parts, tie the cleaning equipment selection to those limits, and put the qualification in writing. That is how a cleaning step stops being a housekeeping task and becomes a controlled input to gearbox reliability. For application review or a line trial, contact PowerDe at sales@powerde.net.
