Emerald Group Publishing Limited
Copyright © 2000, MCB UP Limited
Uncommanded nosewheel steering offset
A Jetstream 31 aircraft was planned to fly a non-scheduled flight from London (Stansted) to Le Bourget. The forecast meteorological conditions for the departure were good. The aircraft was fitted with a four-channel cockpit voice recorder (CVR) only, as allowed by the air navigation order for scale S(i) aircraft, which requires either a CVR or a flight data recorder.
The start was normal except that the left engine was slightly slow to accelerate. This was a known characteristic of this particular engine and was within the published limits. A rolling take-off was initiated using full power. All cockpit indications were normal with both the torque and RPMs at 100 percent and both exhaust gas temperatures (EGTs) at approximately 610°C. The commander, in the left seat, was the handling pilot and initially he used the nosewheel steering to maintain the aircraft on the runway centreline; he later commented that minimal steering inputs were required to track the centreline accurately. At the 70kt call he relinquished control of the nosewheel steering, placed his left hand on the control yoke and called "my stick"; these actions were subsequently verified by the first officer (FO). Analysis of the sound of the aircraft running over the runway centreline lighting at this point showed that the aircraft had achieved a ground speed of 73kt. At approximately 80kt a slight, low frequency vibration was noted by both pilots; however, it was of no particular significance and the commander later stated that he certainly did not consider aborting the take-off. The aircraft then began to turn to the left at a rate that both pilots described as rapid. They later described the motion as being very similar to a simulated engine failure at low speed, although there was no perceived change in engine noise nor was there any cockpit indication or warning of an engine malfunction. Frequency analysis of the area microphone recording on the CVR showed no evidence of engine asymmetry.
As the aircraft yawed to the left the commander applied full right rudder and then right brake but could not contain the heading deviation. As the aircraft was about to leave the runway surface he closed both throttles and once on the grass he selected full reverse thrust on both engines; this symmetric selection of full reverse was verified by analysis of the CVR. The FO notified ATC that they were aborting the take-off as the aircraft left the runway. As the aircraft slowed the commander reverted to use of the nosewheel steering and brought the aircraft to a halt on a heading roughly parallel to the runway heading and about 50 yards displaced from the runway edge.
The parking brake was applied and the commander commenced the emergency shutdown checks while the FO went back into the cabin to organise the passenger evacuation. However, just before the FO left his seat he reported smoke from the engine intake of the right engine. Another aircraft, which had seen the incident, reported to ATC that there was smoke from the left side of the aircraft. The commander shut down both engines using the feather levers; these levers shut both the low pressure and high pressure fuel cocks, feather the propellers, shut the hydraulic cocks and inhibit the engine starting circuits. As the engines were winding down the commander saw flames in the intake of the right engine; he fired both right engine fire extinguishers and told the FO to get the passengers evacuated quickly. The passengers vacated the aircraft through the main door (left rear) without injury. Once the emergency shutdown checks had been completed the crew also left the aircraft. The airfield fire services arrived on site promptly.
It was evident from the marks on the runway that the excursion was caused by a progressive uncommanded nosewheel steering offset to the left. Had the nosewheel been castoring freely, as it should have been at that stage, then no side forces would have been developed by the nose tyres and no marks left by them on the tarmac.
Investigation of the nosewheel steering system
The nose gear steering system is hydraulically controlled via a conventional linear actuator located on the nose gear housing, which pivots the nosewheels about the oleo axis in the conventional manner. The pilot's steering inputs are made using a tiller wheel located on the cockpit left-side control panel. These movements are transferred, via a chain and sprocket system, to the input shaft of the steering control valve assembly. The control valve assembly is mounted in the roof of the nosewheel bay, directly above the nose gear strut, on the steering pivot axis. It comprises a conventional hydraulic spool type control valve, which incorporates a mechanical linkage which sums the pilot's input and the steering-angle negative feedback signals, the latter derived mechanically from the top of the nose gear via a torque link which accommodates the changing orientation of the gear during retraction. The output from the summing link positions the control valve spool, which is of the conventional bobbin type. Although conventional so far as its principles of operation are concerned, the mechanical arrangement of the summing linkage is unusual.
The summing linkage senses relative rotational movements of the input and feedback elements, about the vertical axis. The mechanism is shown schematically in the diagram and consists of the following:
An input shaft, at the top of the unit, driven by a chain sprocket wheel connected to the pilot's control tiller.
A steering angle feedback shaft, at the bottom of the unit. The lower end of this shaft connects via torque link to the top end of the nose wheel gear strut, and rotates with the nose gear as the steering angle changes.
A yoke assembly, integral with the lower end of the input shaft, fitted with roller wheels at each end.
A rocker (also known as the differential), interposed between the input and feedback elements of the linkage. This is pivoted at its lower end on a cross-pin, mounted on the upper end of the feedback shaft.
The outer rim of the rocker has cut-outs on opposite sides which accommodate the rollers of the yoke. One of these cut-outs is a close fit on its roller, the other is larger, providing a clearance between the sides of the cut-out and its associated roller. The lower part of the rocker has a large circumferential groove around its periphery. One end of the control valve input lever engages this groove, such that tilting movement of the rocker on its cross-shaft moves the lever, and positions the control valve spool. The valve spool itself is lightly spring-centred to the null position.
In operation, the pilot's steering inputs turn the input shaft via the sprocket wheel, causing the yoke at its lower end to rotate and the roller to contact the side of the smaller cut-out in the rocker. Because the rocker is not free to rotate, the action of the roller pushing against the side of the cut-out causes the rocker to tilt about the cross-shaft. This tilting movement produces a corresponding movement of the lever engaged in the groove at the bottom of the rocker, which positions the valve spool to direct fluid to the appropriate side of the steering actuator.
The control valve assembly was dismantled and evidence of wear was found at the pointed tips of the rocker return plungers, and at the corresponding contact areas on the underside of the rocker. The wear was more pronounced on the spring cartridge which returned the rocker to the null position from a steer left direction, the resulting excess clearances preventing the rocker from centralising consistently following a displacement in the steer left sense. The affected areas were not lubricated and fretting and wear products were clearly visible.
The worn rocker return spring plunger would have created conditions in which the rocker could have retained a residual steer left offset following an initial displacement in the steer left direction. It was shown during testing that such small offsets of the rocker could overcome the very light centring spring of the valve spool, resulting in a small valve offset and associated fluid flow to the steering actuator. It would appear that a residual offset occurred for some reason during the take-off, and that offset was effectively masked by the slack in the feedback linkage. As a consequence, the normal corrective action of the feedback mechanism was inhibited resulting in a progressive displacement of the nosewheels.
It was apparent, both from the initial position of the nose tyre marks on the runway and from the CVR recording, that the aircraft was accurately tracking the centreline and that the nose wheels were thumping the runway centreline lights as it did so. Because of the twin nosewheel configuration, this would have generated a succession of torsional shock loads, about the steering axis, which would have been fed back up through the nose leg and into the steering feedback linkage. During the initial stages of take-off, the tiller was being actively used by the pilot to hold the aircraft straight and so the fault condition would have been unlikely to arise. However, after tiller release, the nose leg would have reverted to castering mode, and it is probable that the shock loads fed back to the control valve via the feedback link were sufficient to displace the valve spool slightly, into the dead band just to the left side of neutral, while remaining within the regime of slackness present in the feedback system. This is the most probable explanation for the progressive uncommanded steer left movement at the actuator.
As a result of these and other actions, two service bulletins were issued:
(original issue date 28 October 1998) introduced measures to inspect in situ for wear in the feedback and input linkages, with appropriate rectification actions.
(original issue date 28 October 1998) introduced a one time requirement for removal of the steering control valve and return to an approved agency for overhaul, and subsequent overhaul at 10,000 hours intervals.
These service bulletins have been subject to subsequent review and amendment in the light of ongoing experience, and revisions incorporated into the maintenance schedule and appropriate manuals to bring these into line with the service bulletin requirements.
AAIB Bulletin 1/2000