Quieting counter-rotating aircraft gearboxes with mesh phasing
When two rotors spin in opposite directions inside a power reduction gearbox for a modern aero-engine, the tooth engagements at each stage march to their own clock. Without deliberate alignment, the result is a chorus of vibration that wears bearings, fatigues housings, and bleeds efficiency from a drivetrain built to shave every gram. Multi-stage counter-rotating arrangements make this worse, because each stage adds its own mesh frequency to the acoustic and mechanical spectrum.
Mesh phasing is the engineering practice of choosing the angular position of each gear so that excitation peaks line up against each other in a way that cancels out rather than compounds. In a planetary or epicyclic stage, this often means offsetting the planet gears relative to one another around the carrier. Across multiple stages of a counter-rotating gearbox the choice grows more involved, because rotation directions, speeds, and tooth counts all interact.
Australia sits at an interesting vantage point for this kind of work. Melbourne's Fishermans Bend precinct, with deep roots in aero-component manufacture, has long hosted suppliers building power-turbine casings and accessory-drive housings. Adelaide's Lot Fourteen precinct and the wider DefenceSA ecosystem connect that manufacturing to programmes such as the F-35 strike fighter and the P-8A maritime patrol aircraft, both of which depend on advanced gear systems. Under Civil Aviation Safety Authority airworthiness rules, vibration and bearing-load reductions feed directly into the fatigue and endurance substantiation Australian-modification organisations must demonstrate, making the case for phasing as much a certification matter as a mechanical one.
The rest of this article walks through how phasing works, why it is so effective at cutting vibration in counter-rotating stacks, and how design-of-experiments, simulation, and physical testing fit together in the development loop. Practical guidance for engineers weighing whether to introduce phasing into their next gearbox architecture appears at the end.
The mechanics of counter-rotation and mesh timing
A counter-rotating gearbox is configured so that two output sections of the drivetrain turn in opposite senses. In a typical power reduction gearbox, the input shaft drives a first stage one way and a second stage the other, to feed two spools of a high-bypass turbofan or a contrarotating open rotor. Each meshing pair contributes an excitation frequency equal to the tooth count multiplied by the rotational speed of that gear, plus higher harmonics.
The amplitude of mesh excitation depends on transmitted torque, contact ratio, profile modification, lead crown, and manufacturing variation. Transmission error, the small deviation between theoretical and actual motion at the pitch point, is the main driver of dynamic mesh forces. When two stages run with arbitrary tooth phasing, their transmission-error harmonics arrive at the housing at slightly different times, producing beat patterns, sidebands, and broadband excitation that the rotor-dynamic model has to absorb.
In a single-stage epicyclic gearbox with three or four planets, the canonical choice is to space the planets evenly around the carrier so that their static load sharing is symmetric. Even spacing does not necessarily minimise vibration, however, because each planet-to-sun and planet-to-ring contact generates a force vector whose direction rotates with the carrier. Phase offset between planets can be tuned so the resultant vector stays closer to the carrier axis of symmetry, reducing pendulum-mode excitation and bearing whirl.
Why phasing cancels vibration across counter-rotating stages
The cancellation principle is straightforward in the frequency domain. When two stages produce mesh excitations at the same frequency but slightly different phases, the combined force is the vector sum of the individual forces. Setting the relative phase near 180 degrees for the dominant harmonic can drop the resulting vibration by a factor of three or more.
In a counter-rotating arrangement, the stages turn in opposite directions, so fundamental mesh frequencies rarely match even when tooth counts and shaft speeds suggest they should. Phasing therefore targets harmonics and structural mode shapes, not just first-order mesh pulses. Engineers build a Campbell diagram that overlays shaft-order excitations against housing bending modes and bearing-support stiffness, and phasing is selected so that the largest mesh harmonics land in valleys of structural receptance rather than on its peaks.
A second mechanism is what rotor dynamicists call load-sharing homogenisation. Even when two planets carry slightly different static loads due to manufacturing variation, deliberate phase offsets across a multi-stage system can spread the dynamic loading in time so that no single planet-to-sun or planet-to-ring pair sees the worst combination at the same instant. The effect shows up cleanly in bearing reaction-force measurements and in acoustic intensity scans on the housing. A habit of literature review reading engineers can build into a weekly routine tends to surface new phasing heuristics from universities and OEMs before they cross into the mainstream press.
| Phasing strategy | Mesh harmonic reduction | Main application | Sensitivity to torque |
|---|---|---|---|
| Zero-phase (factory default) | Baseline | Single-stage units with low dynamic demand | Low |
| Equal planets, stage phase offset | 20-40% on dominant harmonic | Two-stage counter-rotating units | Moderate |
| Alternating opposite-stage phase | 30-55% across first three harmonics | High-speed PGBs for turbofans | Moderate to high |
| Computed modal-phasing (DoE-tuned) | 50-70% across broadband | High-power-density military and eVTOL | High, requires sensing |
Four common strategies are summarised here, comparing them on mesh harmonic reduction, application, and torque sensitivity. The values are typical of the geometries found in modern power reduction gearboxes rather than universal, and they show how engineering returns grow as phasing is matched more carefully to the structural modes and operating envelope. Computed modal phasing, tuned through design of experiments and validated on rig, is the path the OPTIMIZE programme has chosen to pursue, and the broader objectives of the project are framed around exactly this kind of broadband suppression.
Simulation and design of experiments in phasing studies
Picking a phase angle by intuition is rarely successful in a multi-stage counter-rotating unit, so the industry has moved heavily on simulation and statistical experimentation. A modern phasing study starts with a finite-element model of the gear blank, a housing model that captures the bearing webs and the bolt-up region, and a rigid-body or flexible-body dynamic model of the gear train. Transmission error from the loaded contact analysis feeds the dynamic solver, which then yields time-domain bearing forces and housing surface velocities.
The number of variables quickly grows: planet phase angles in each stage, profile-relief lead crown, lead-lag of the input pinion, mounting stiffness, and lubrication regimes. A full factorial on every variable is impractical, even on a high-performance cluster. Design of experiments methods such as Latin hypercube sampling, optimal space-filling designs, and Taguchi arrays let engineers interrogate dozens of factors with hundreds of runs rather than millions, identifying the few that dominate the response. Companion work on housing strain, discussed in a recent article on using a design of experiments on gearbox mounting locations to reduce housing strain, shows how the same toolset supports housing-side optimisation in parallel.
Sensitivity analysis typically narrows the field to two or three variables that explain most of the variance. Gradient-based optimisation can then drive the dominant mesh harmonic to a local minimum. The output is rarely a single number but a Pareto frontier that balances vibration, weight, efficiency, and manufacturing cost. Australian researchers at CSIRO's Data61 and at university labs in Adelaide and Brisbane have been steadily contributing to the open literature on design-of-experiments workflows for aerospace gear systems, keeping the local skill base close to the leading edge.
Tolerancing, lubrication, and operating speed effects
A phasing strategy that works perfectly in the computer can lose its edge at the production line if tolerance scatter eats into the phase relationship. Tooth thickness variation, helix slope errors, and runout all shift the effective mesh phase of every gear. When a fleet of gearboxes is built from thousands of components, the distribution of these errors matters more than the nominal design. Tolerance analysis tools such as RSS, Monte Carlo, and 3DCS are now standard parts of the gearbox engineering toolset in Australian overhaul shops, particularly under CASA rules that require traceable statistical evidence of compliance for any modification affecting fatigue-critical components.
Lubrication adds another layer. Oil film thickness influences effective centre distance, which shifts the phase of mesh excitation. At the high operating speeds of a modern aero-engine gearbox, where pitch line velocities can exceed 100 metres per second, elastohydrodynamic films become both thinner and thicker across different regimes. Jet-spray lubrication can introduce local thermal gradients that warp the housing and again alter phase. Phasing studies must therefore be coupled with thermal and lubrication models so the chosen phase angle survives a full takeoff, climb, cruise, and descent cycle.
Operating speed varies across the mission profile, so optimum phasing is speed-dependent. A phase angle that suppresses the second mesh harmonic at cruise can leave the third harmonic untouched at shaft redline. For engine programmes targeting both high-speed cruise and demanding transient ratings, the phasing search has to sweep across the full envelope. The rig-side counterpart of this analysis is described further on, but the simulation already tells engineers that a robust phasing solution flattens the response across a range of speeds rather than minimising it at a single point.
Validation through rig testing and flight-relevant conditions
Simulation work finds the candidate phase angles, but they have to survive the rig. A purpose-built test stand, instrumented with accelerometers on the housing, bearing reaction sensors, acoustic microphones, and high-speed torque telemetry, is the judge. Rig tests measure vibration amplitude, bearing temperature, lubricant film pressure, and gear tooth-root strain, giving engineers a multi-channel picture of whether the phasing improvement transfers to the parts that matter in service.
A representative test campaign starts with a checkout run on a baseline gearbox with factory-default phasing, then steps through the candidate configurations one at a time. Measurement-while-running sweeps through takeoff power, cruise power, and a series of steady-state conditions so that sensitivity to speed is captured. Back-to-back comparisons between configurations show the vibration reduction directly, without any assumption about the underlying model. Australian test facilities such as Defence Science and Technology Group laboratories and private rigs in Melbourne and Brisbane routinely run this kind of campaign under AS 1100-compliant drawing standards and CASA audit expectations.
Endurance runs follow the success cases. A phasing arrangement that delivers 60 percent vibration reduction at 10 hours but causes premature pitting at 500 hours is not a success, regardless of how it looks on the spectrum analyser. The OPTIMIZE programme and partner laboratories therefore test endurance and efficiency together, and they correlate rig results with single-tooth bending fatigue coupons and rolling-contact fatigue coupons run on subsidiary rigs. The outcome is a phasing choice that has survived statistical scrutiny, simulation, and physical service loading, ready for flight-clearance substantiation.
Guidance for engineers considering phasing
When evaluating whether to introduce mesh phasing into a new counter-rotating gearbox design or a retrofittable upgrade, the following points are worth working through methodically.
- Begin with a baseline modal and rotor-dynamic model that includes the housing, bearing supports, and gear train dynamics, then validate it on the existing rig before changing any phase angles.
- Use design of experiments to sweep planet phase, profile modification, lead crown, and housing stiffness together, rather than optimising each in isolation.
- Couple the dynamic model with thermal and elastohydrodynamic lubrication models so the optimum remains valid across the full mission profile.
- Run tolerance analysis with Monte Carlo methods, paying particular attention to tooth thickness, helix slope, and runout, because these directly influence effective mesh phase.
- Specify at least two candidate phasing configurations for rig testing, not just the optimum from simulation, to bound the sensitivity and confirm robustness.
- Plan an endurance run of at least several hundred hours under representative load before declaring success, because reduced vibration is meaningless if pitting or spalling accelerates elsewhere.
- Document the statistical rationale for the chosen phase so that the certification case, including CASA airworthiness substantiation, can withstand independent review.
A growing set of references helps engineers stay current with the state of the art, including publications from academic teams in Adelaide and Melbourne that frequently collaborate with European partners through the Clean Sky family of programmes. For practitioners who want to broaden their reading across both technical and adjacent fields, a few industry reading hubs offer starting points for stepping outside pure engineering literature into the wider systems-engineering conversation.
If you are working on a counter-rotating power reduction gearbox, the OPTIMIZE programme invites collaboration across simulation, testing, and design-of-experiments. Reach out through the project site to explore how mesh phasing can be tuned to your architecture, or to schedule time on the project's test infrastructure.