Bearing raceway roughness and high-speed gearbox friction
The effect of bearing raceway roughness on friction torque in a high-speed gearbox is a micron-scale issue with propulsion-scale consequences. A change in the finish of a raceway can alter lubricant entrainment, asperity contact, and the heat returned to the gearbox housing. In a geared aircraft engine, lower bearing drag supports mechanical efficiency, thermal margin, and usable power density. It also reduces the cooling burden on the oil system, which matters when the gearbox must operate reliably during take-off, climb, or rapid changes in power.
Roughness is not adequately represented by one number. Ra, the common average-roughness measure, can hide tall peaks, steep slopes, and directional lay that may dominate local contact behaviour. Engineers should also consider Rz and related height parameters, roughness slope, and the direction of grinding or finishing marks. Waviness, roundness, alignment, and raceway geometry need separate attention because a nominally smooth surface can still produce uneven loading if it is wavy or installed with the wrong preload.
A credible research programme therefore connects metrology to a measurable loss. The OPTIMIZE Project applies design-of-experiments methods, simulation, tolerance analysis, and physical testing to the design of power-reduction gearboxes for geared aircraft engines. Candidate surface finishes can be compared under controlled speed, preload, lubricant, and temperature conditions, while models estimate how contact mechanics and lubricant-film changes contribute to drag. The approach also tests whether a proposed improvement remains worthwhile after manufacturing variation and long-duration wear are considered.
Australian operating conditions make that translation essential. Engineering and aerospace manufacturing activity around Brisbane, Melbourne, and Adelaide serves aircraft operating from hot, dusty continental airports and humid coastal environments near Sydney. CASA airworthiness requirements place a premium on traceable specifications and repeatable acceptance tests. A bearing that performs well in a laboratory-clean environment may still encounter contamination, heat, or maintenance effects in service, so local conditions belong in the test matrix rather than being added after the experiment.
| Surface condition | Contact and lubrication behaviour | Friction-torque expectation | Engineering response |
|---|---|---|---|
| Controlled polished finish | Fewer asperity interactions; stable film when load and viscosity are suitable | Often lower baseline drag and run-to-run spread, but not zero | Set Ra plus slope, lay, and a sampling rule |
| Directional lay | Rolling marks can guide lubricant along the contact path | The result depends strongly on rolling direction and alignment | Specify direction and inspect after handling |
| Coarse or irregular texture | Peaks carry local load and disturb the lubricant film | Higher mean torque, heat, and variability are likely | Limit height and slope; validate at operating speed |
| Damage or debris | Spalling, contamination, and edge loading dominate | Nonlinear torque rise and vibration may occur | Reject, inspect, and protect the lubrication system |
Why raceway roughness changes friction torque
Within a rolling bearing, the rolling element and raceway meet in a small Hertzian contact area, not a broad uniform surface. Normal load, preload, and local geometry determine how much force is transmitted; lubricant film and microscopic asperities determine how it is shared. As the elements roll and slip through contact, tangential resistance produces friction torque. The simple relationship (T_f \approx \mu F_n r) links friction coefficient, normal load, and effective radius, but the actual coefficient changes with contact conditions and surface texture.
At high speed, film formation, viscous shear, centrifugal effects, and thermal expansion compete with roughness effects. When a full lubricant film separates the surfaces, roughness may have only a modest effect on total drag; in mixed or boundary lubrication, asperity interaction can become decisive. The relationship is not necessarily linear. A smoother finish may reduce ploughing, yet an overly smooth or incorrectly finished surface can alter lubricant retention or run-in behaviour. Preload, viscosity, alignment, cage drag, and contamination must be held or measured before attributing a torque change to raceway roughness alone. In hyperstatic bearings, where preload exceeds the changing external load, small changes in conformity can magnify sensitivity to texture; in other designs, the same finish may have little effect.
From surface measurement to contact modelling
Reliable analysis starts with how the finish is measured, not with the finish designation on a drawing. A stylus profilometer can provide repeatable Ra and profile data, while optical or atomic-force methods can characterise very fine features. Measurements should cover several positions around the raceway and more than one rolling path, with calibration standards, filtering choices, sampling length, and trace location recorded. For instrument selection and calibration, precision test equipment offers a useful reference point before a campaign begins.
The resulting data can feed contact models rather than sit in a separate quality report. Asperity models estimate how compliant peaks carry load and create shear; elastohydrodynamic models estimate film thickness, pressure, and entrainment speed. Coupled simulation translates local changes into bearing drag and gearbox heat, then tests sensitivity to preload, clearance, oil viscosity, and material properties. Design-of-experiments methods select a compact set of roughness levels and operating points, while tolerance analysis identifies which characteristics matter most when real components depart from nominal. Physical testing anchors the model and exposes nonlinear effects that a single baseline comparison could miss.
Designing a high-speed test campaign
A representative campaign must define what friction is being measured. A bearing-on-shaft arrangement isolates rolling-element losses, whereas a complete gearbox adds gear-mesh, seal, and windage contributions. The rig needs a high-speed drive, torque transducer, temperature monitoring, controlled lubricant delivery, and a clear record of speed and preload. The project members area provides access to project material that can support this planning, including the objectives and methodology behind the experimental work.
A factorial design is more informative than changing one finish at a time. Roughness can be a factor with deliberately manufactured levels, while speed, preload, lubricant viscosity, inlet temperature, and lubricant pressure are controlled or deliberately varied. Randomised runs and repeated baselines separate an effect caused by surface texture from drift in the rig, and intermediate roughness values can reveal a threshold or mixed-lubrication transition. Imported torque dynamometers and profilometers also require Australian Customs clearance, GST treatment, and correct commercial paperwork, so procurement lead time belongs in the experimental plan.
Measurement priorities
- Record roughness, waviness, and lay across the same bearing zones
- Use calibrated instruments and a defined filtering protocol
- Compare several circumferential positions and rolling directions
- Link every surface result to torque, temperature, and wear data
Manufacturing and validation priorities
- Produce and inspect controlled roughness levels on each candidate finish
- Check specimen geometry, roundness, alignment, and preload before testing
- Repeat baseline runs to quantify rig repeatability and drift
- Preserve parts and measurements for traceability after each test
Reading results in Australian operating conditions
Results become more useful when they are read against Australian service conditions. Aircraft may operate from hot, dusty inland airfields, while coastal operations expose filters and seals to salt-laden air and humidity. Dust can raise contact stress and break down a lubricant film; temperature alters viscosity and preload; contaminated oil can turn a small surface feature into a wear problem. The relevant question is not whether a laboratory raceway is smooth, but whether its friction, heat, and durability remain controlled where the gearbox operates.
Windage must also be separated from bearing friction before a result is used in design. Airflow over gears and shafts changes with ambient pressure, and that drag can be mistaken for a change in rolling contact. The project's windage test rig method explains how ambient and vacuum conditions can quantify the difference. Pairing those measurements with bearing temperature, oil pressure, and rotational speed gives a defensible loss budget. CASA expectations and Australian operators' need for predictable component life then make the findings directly relevant to airworthiness and maintenance decisions.
Designing with tolerances, lubrication and manufacturability
A surface-finish requirement should be a tolerance system rather than a single target. The drawing may need to specify roughness, directional lay, waviness, roundness, alignment, and hardness, while tolerance analysis checks whether these characteristics can coexist within the available preload and manufacturing process. In a hyperstatic arrangement, a small change in raceway conformity can alter contact pressure and make friction torque more sensitive to texture. Acceptance criteria should identify the measurement method and sampling rule, or two suppliers may deliver parts that look equivalent but behave differently.
Lubrication is the other half of the design decision. Oil viscosity, base-oil chemistry, additive package, temperature, and flow rate determine whether the raceway runs in a robust film, a mixed regime, or a boundary condition, so lubricant choice and surface finish should be tested together. Seals, filtration, and contamination control determine whether a measured gain survives service. The project's technical documentation can support a traceable test and inspection record, linking the selected finish to torque and durability evidence while local suppliers and NATA-accredited calibration laboratories provide a practical verification chain.
Turning surface data into a better gearbox
The strongest outcome is an integrated design loop. Surface data from the finished bearing should feed the contact model, torque budget, and tolerance analysis; test results should then update the manufacturing specification. Engineers can choose the finish that balances friction reduction, thermal margin, fatigue life, weight, and power density rather than pursuing the smoothest surface at any cost. For Australian programmes, the evidence package should also address local procurement, inspection capability, airworthiness implications, and the conditions encountered by operators, with project videos and documented methods making the reasoning accessible across design, manufacturing, and maintenance teams.
Use the OPTIMIZE Project methodology to turn the next gearbox requirement into an evidence plan: define representative roughness levels, specify the speed, preload, lubricant, and temperature matrix, and build a torque-and-wear measurement plan around those variables. Review the available project guidance, then arrange the first controlled comparison with the people who will manufacture and maintain the hardware. Carry only a validated finish into the qualified design.