Incident Overview
A scheduled domestic passenger flight of a Boeing 737-322 was descending for a visual approach to the destination airport about 30 minutes after departure. The aircraft, which had reached a maximum altitude of approximately 27,000 feet with an outside air temperature of -36°C, encountered wake turbulence while descending through 9,300 feet mean sea level. The airplane began a correction back to the right while continuing to bounce, rapidly banking 10 degrees. The captain disengaged the autopilot and autothrottles, applying left aileron and left rudder to oppose the roll; the first officer subsequently assisted with control inputs.
Digital flight data recorder (DFDR) data showed a right 26-degree roll, a left 45-degree control wheel deflection followed by right deflections, and rudder pedal deflections reaching a maximum of left 6.5 degrees. The rudder pedal then remained deflected left between 2.0 and 0.25 degrees until landing. The first officer reported that hydraulic quantity and pressure gauges for systems A and B indicated normal values with no changes during uncommanded rudder checklist items. During the landing roll at about 100 knots indicated airspeed, the captain heard a banging noise and experienced a full right rudder pedal deflection. The cockpit voice recorder recording was not retained by the operator.
Cockpit Inspection
Inspection of the cockpit nose area forward of the main landing gear revealed a wood-handled wire brush. The first officer's pedal arms contained scuff marks and gouge marks that penetrated the paint and metal. A borescope inspection showed displaced white paint on the structure aft of the pedal arms, though no dents were found on the vertical structure. The United Airlines party representative stated that work related to an airworthiness directive had been performed weeks prior to the incident.
The brush, a new brush, and the first officer's pedal arms were sent to the National Transportation Safety Board's Materials Laboratory. Examination of the pedal arms and brush revealed light score marks on the boss consistent with tool usage. The right pedal arm had one score mark on the rear edge of the boss, circumferential for approximately 90 degrees of arc centered at the horizontal line, more significant above the horizontal and tapering below. The brush handle showed no crush damage. Functional testing and visual examination of the rudder power control unit revealed no anomalies.
Examination and Simulation
Estimates of maximum rudder forces ranged from 246 to 424 pounds depending on pedal position and seat distance. Similar brushes were tested: when placed in obstructive positions within the pedal arms, a pilot applied about 15-20 pounds of force to fracture them; when loaded lengthwise, brittleness varied, with breakage occurring at 80 pounds or less.
A flight training simulator at United Airlines' Flight Training Facility was back-driven using FDR data from the incident flight with both crew members aboard. The simulator's rudder pedal area had the same platform/shelf structure as the incident airplane. With the pedals set to the first officer's reported adjustment, a brush was inserted while applying left rudder pedal pressure to hold it in place (the simulator's resting platform was shorter than the actual airplane). The strongest holding position was with the wire bristles facing downward and toward the rudder control arms, the convex portion of the handle upward, and the other end against the vertical wall of the pedal box. No further pressure was applied due to concern for damaging the simulator.
A search of the Federal Aviation Administration's service difficulty reporting system for "Boeing Rudder Controls" from January 1, 1997 to February 1, 2002 identified three occurrences: one where no fault was found and rudder hydraulic components were replaced; a second with no defects after engineering callouts; and a third where the power control unit was replaced.