On July 25, 2006, at Spanish Fork, Utah, the prototype of the Spectrum Model 33 business jet lifted off on a test flight. The 7,300-pound all-composite airplane, powered by two Williams FJ33 turbofans of 1,562 pounds of thrust each, had previously logged 44 hours in the course of 46 test flights before spending seven weeks on the ground for modifications to the main landing gear and the surrounding structure and systems.
Immediately after breaking ground, the airplane rolled to the right. The wing struck the ground and the airplane crashed, killing both pilots. Subsequent examination of the wreckage revealed that the aileron controls had been inadvertently installed in such a way as to reverse their operation.
The flight control system was mechanical, consisting of pushrods, bellcranks and torque tubes.
Such systems are uncommon on jets, because it is difficult to ensure reasonably low control forces — especially with sidesticks, which this airplane had — at high speeds. They have the advantages, however, of simplicity, low cost and light weight, and, once installed, require little maintenance.
The major maintenance, in this case, was being performed to increase the stiffness of the main landing gear. The aileron linkage got involved only because some of its elements occupied the same space in the fuselage behind the pressure bulkhead as the landing gear retraction system did.
When the modified landing gear parts were installed, it was found that they interfered with a bracket supporting one of the aileron torque tubes. A new support was designed and fabricated.
To slide the torque tube into the new support, mechanics had to remove a bellcrank. This bellcrank was secured to the torque tube by a single tapered pin.
The mechanic who did the work believed that the use of a tapered pin, which meant that holes on opposite sides of the bellcrank were of different sizes, made it impossible to install the bellcrank incorrectly. But this was not the case.
If the bellcrank was slid onto the tube facing the wrong way and then rotated 180 degrees around the tube, the tapered holes would again line up. That’s what he unwittingly did.
When the mechanic installed the torque tube in the airplane, he found that the roll-control system bound up, with hardly any movement.
Inspecting the situation, he saw what he believed to be his error and turned the torque tube over, so that the bellcrank, which had been tilting upward, was now tilting downward.
After adjusting the length of a pushrod to suit the new position of the bellcrank, he found that the ailerons now worked, except for a little rubbing that was relieved by adding washers under a rod end. The pilots tried the new system and accepted it.
No one noticed that when a side-stick was moved to the left, the left aileron moved downward and the right aileron upward.
The National Transportation Safety Board, in its finding of probable cause, noted that because this was a proof-of-concept prototype there was no maintenance documentation to guide the mechanic. The mechanic, in turn, believed that the bellcrank could be installed in only one way.
Indeed, in the olden days, before the invention of highly precise computer-controlled machining equipment, that would very likely have been true; it would have been very difficult, and furthermore pointless, to align the tapered pin so perfectly with the center of the tube that it could be assembled in two different ways.
His belief prevented him from recognizing that the various interferences that arose afterwards, and the need to change the length of a pushrod, were hints that something was wrong with the way the parts had been assembled in the first place. It was extraordinarily bad luck that an alternative way of combining them was even possible, and that it resulted in reversing operation of the ailerons.
The Board faulted the maintenance personnel and flight crew for failing to verify that the controls were operating in the desired directions. Checklists said to “check” the controls, but did not say what to check for. In general, pilots look out the windows to verify that whatever controls they can see are moving, but do not stop to inquire whether they’re moving correctly. After all, they always are.
This accident may seem unique, or nearly so, and therefore to have little bearing on flight operations generally. But it is not unique. Other instances of control reversal, typically because of incorrect cable hookup, have occurred. Significantly, however, a couple of recent instances in which crews lived to tell the tale reveal that the danger is not confined to mechanical-control systems.
In 2001, a Lufthansa Airbus A320, taking off from Frankfurt on a revenue flight after maintenance had been performed on the captain’s sidestick wiring, encountered wake turbulence, which banked the airplane slightly to the left. The captain, who was the pilot flying, responded with right sidestick, whereupon the airplane rolled farther to the left. The first officer somehow recognized the problem, took over control priority, and leveled the wings after the left wingtip had come within 20 inches of the ground.
The A320 is a fly-by-wire airplane. The cause of the problem was an accidental reversal of the polarities of four wires during reassembly of an electrical connector in which a damaged pin had been replaced.
The repair involved several shifts of technicians over several days; the final check involved verifying control response to stick movement, but the mechanic performing the check used the first officer’s sidestick.
It’s not clear, however, that even if the captain’s sidestick had been used the technician would have detected the incongruity between the stick movement and the control reaction displayed on the electronic control monitoring system.
Maintenance techs aren’t required to be pilots or aerodynamicists.
The control check was performed by the flight crew during taxi out, but, as most flight crews would, they noted only control movement, not the correlation of the control movement direction with the stick movement.
In November 2018, an Air Astana Embraer 190 departing Lisbon after maintenance experienced aileron reversal in instrument conditions.
According to one report, the airplane performed several barrel rolls, after one of which it found itself at 4,000 feet pointed straight down. Eventually, the crew, displaying remarkable airmanship, got the airplane under control, discovered that it would fly normally on autopilot, and managed to make a safe landing.
The common element among these events is that the flights occurred immediately after maintenance. If, in rare instances, mechanical assemblies can be put together incorrectly, the electronic assemblies of fly-by-wire airplanes, which are infinitely more complex, offer new opportunities for error.
Fortunately, pilots have it in their power to detect errors on the ground rather than discover them in the air. It is merely necessary to perform the routine control checks with awareness not only of control movement, but of the direction of the movement.
When a roll-left command is given, the left aileron should go up, the right one down. A pitch-up command should raise the trailing edge of the elevator on a conventional airplane and lower it on a canard. These facts seem almost too elementary to mention; yet few pilots actually attend to them. All should.
