AAIB Bulletin: 2/2009

distorted on its aft face, and there was ‘tearing’ of both tube sidewalls where the pivot tube had been secured. Damaged water rudder hinge post Microscopic examination revealed that the upper and lower failure surfaces, together with the ‘tears’ in the post sidewall, exhibited characteristics of a failure due to overload. The circumferential welds securing the pivot tube to the rudder post had not completely penetrated the post sidewall, and had also failed in overload, Figure 4. Both of the rudder post sidewalls in the region of the circumferential welds had been distorted by the application of a torsional load in a clockwise sense (viewed from the rear). There was no evidence of corrosion or crack progression by a fatigue mechanism on any of the fracture faces. When operating from water, the water rudder pivot is below the waterline which together with the presence of a liberal 135 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-MDJE EW/G2008/05/29 Figure 3 Damaged water rudder hinge post coating of protective grease, may have prevented the damage to the post being observed during a pre-flight inspection on the water. Analysis The operator has highlighted to the manufacturer the disparity between the cable tension requirements detailed in the float installation drawing and the manufacturer’s service manual. The manufacturuer is reviewing the situation and has undertaken to issue clarification of the cable tension requirements. 136 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-MDJE EW/G2008/05/29 Figure 4 Rudder post weld details The deformation of the rudder mounting post indicated that the rudder pivot tube had been pulled aft, causing the aft wall of the rudder post to fail where the pivot tube was attached. The ‘tears’ in the post sidewalls illustrate that, despite the welds securing the pivot tube having only partial material penetration, they were able to transmit sufficient load to damage the rudder post sidewalls without failing. The damage observed was consistent with a force being applied to the lower portion of the rudder from the rear whilst deployed. This caused the rudder to ‘pivot’ about the lower mechanical stop, producing the rearward bending of the post where the pivot tube was attached. The force was of sufficient magnitude to cause the failure in overload of the two longitudinal welds securing the pivot tube. The distortion of the post sidewalls close to the circumferential welds showed that a torsional load had been applied through the pivot tube, in a clockwise sense (aft looking forward), after the failure of the upper and lower longitudinal welds which resulted in the failure of the sidewalls. The subsequent failure of the remaining circumferential welds allowed the rudder to separate from the aircraft. It could not be determined when the damage to the 137 © Crown copyright 2009 AAIB Bulletin: 2/2009 G-MDJE EW/G2008/05/29 rudder post had occurred. The bending of the steering input bar indicated that, either the rudder had been restrained against a right rudder input from the cockpit or, more probably, given the use of cables in the water rudder system, an external force being applied to the rudder, forcing it to the left. The damage to the rear bulkhead of the float was indicative of ‘overtravel’ of the water rudder. It is considered possible that this could be caused by the water rudder cables being incorrectly rigged. However, the aircraft records made available during the investigation showed no evidence of this prior to the incident. The damage to the rudder mounting post was inconsistent with it being damaged as a result of hitting the rear bulkhead of the float. Conclusions Given that the rudder became detached from the aircraft in-flight, the force applied to the rudder, although sufficient to cause significant damage to the mounting post, did not result in the immediate separation of the rudder. Whilst the point at which the rudder became damaged was not identified, the lack of evidence of crack progression through fatigue suggests that the damage had not been present on the aircraft for a prolonged period. Given the rudder’s location, the damage to the post should have been visible during an ‘out of water’ pre-flight inspection but unlikely to have been seen during an inspection whilst water-borne. All the damage observed on the rudder mounting post was consistent with the rudder being struck from behind, with some force, whilst deployed. It is considered that this was more likely to have occurred whilst the aircraft was being manoeuvred on the water, when it may have ‘reversed’ into a submerged object, rather than being struck from behind by a moving object.