Emergency AD targets Eclipse thrust levers

 - July 7, 2008, 6:10 AM

The pilot of an Eclipse 500 pushed the throttles forward during a wind-shear encounter at Chicago Midway Airport last month with enough force to result in an “eng control fail” crew alerting system message followed by a maximum uncontrolled thrust condition on both of the airplane’s Pratt & Whitney Canada PW610F turbofans. Unable to slow the airplane for landing, the pilot elected to shut down one engine for the subsequent landing attempt.  When he did, the thrust on the good engine dropped to idle and became unresponsive to thrust lever inputs, resulting in a hard landing that blew out both of the airplane’s main tires.

None of the four people aboard Eclipse 500 N612KB were hurt in the landing, but the June 5 incident prompted an emergency AD on June 12 requiring inspections of the thrust levers in more than 200 in-service airplanes to ensure there is no indication of binding or “unusual noises,” such as grinding or scraping that might suggest a defective throttle quadrant.

Most of the airplanes in the fleet had completed the required checks at press time with no anomalies uncovered, according to manufacturer Eclipse Aviation. Although the investigation into what caused the out-of-control thrust event is continuing, Eclipse and the FAA advised operators last month of the possibility of exceeding the maximum range of throttle travel resulting in a loss of position signal to the engines’ fadec. This would cause the engine thrust setting to remain in the last known position, in this case maximum.

The FAA and Eclipse were scheduled to participate in a teardown of the throttle quadrant from the incident airplane late last month as part of the incident investigation with the NTSB. Eclipse CEO Vern Raburn estimated that 50 to 60 pounds of force put N612KB’s thrust levers into the out-of-range condition, although testing by Eclipse indicated that a force of 30 pounds against the forward throttle stops could cause the condition to occur.

Raburn said the failed system in N612KB appears to have performed as originally intended. The fadec software, he explained, is designed to “fail fixed.” In other words, rather than spooling the engine back to idle in the event of a loss of engine control, the system leaves the power at the last known setting. “We do know that if you push on the thrust levers hard enough, you can cause this fault condition to occur,” Raburn said, “and we’re trying to get to the bottom of exactly why that is.”

Asked if the issue with the Eclipse 500 throttle quadrant was physical in nature or limited to a fadec software issue, Raburn said more testing was needed before making a determination. “We don’t know for sure yet,” he said. “The answer could be both. There could be a physical buildup of tolerances, which then allows the throttle quadrant to send a signal to the fadec that says, ‘I’m past my forward limit,’ or it could be that the forward limit is set adequately but there’s not enough margin. We’re in the process of going through that analysis” and will try to find a root fix for whatever the cause turns out to be.

Possible remedies could include rewriting the engine fadec software code to prevent such an incident from occurring in the future or replacing the thrust-lever quadrant with a different design. One possible remedy would be to have the fadec attempt to determine multiple times over the course of several seconds whether a failure really had occurred before entering its failure mode, Raburn said. He added Eclipse is continuing to build and deliver airplanes while the investigation into the incident continues. No Eclipse 500s at press time had been grounded as a result of the emergency AD, he said.

An equally important question relating to this incident is why the power on the engine that was left running during N612KB’s second approach to Chicago Midway reduced to idle when the first engine was shut down. Raburn said the system performed exactly as designed, but he acknowledged a flaw in the design that wasn’t anticipated during the Eclipse 500’s test program. “When [the pilot of N612KB] shut down one engine, that cleared the control fault,” Raburn explained. “When the fadec cleared the control fault, it inferred the thrust lever position on the good engine. In other words, the engine he shut down suddenly became a good engine in the sense of absence of fault.”

The pilot by that point had pulled both engines back to idle. When the system reset itself it rolled the power back to idle, yet throttle inputs by the pilot still had no effect on power setting. “We screwed up on that,” Raburn said. “We didn’t anticipate–and in the certification testing, the FAA didn’t anticipate–this dual-fault problem, and we didn’t have an adequate procedure” to deal with such an occurrence.

Now there is a procedure, Raburn said. In addition to the one-time throttle quadrant test required by the FAA’s emergency AD, the directive also includes revised aircraft flight manual procedures that enable the pilot to regain throttle functionality
in the event that a similar incident occurs in the future. “But it’s really easy to avoid this from happening,” Raburn said. “Don’t slam the throttle levers forward.”