Fly-By-Wire Helicopter, Military Pilot Had ‘Conflict’ Before Fatal Crash


Canadian military authorities say the pilot of a Sikorsky CH-148 Cyclone anti-submarine helicopter was apparently overruled by the aircraft’s fly-by-wire flight director and autopilot just before the aircraft crashed in the ocean off Greece on April 30. Preliminary assessment of the flight data recorder information recovered from the wreck of the helicopter, a military version of the S-92, showed the pilot had a “conflict” with the aircraft’s computer and the computer won. He and five other crew members were killed. Lt.-Gen. Alain Pelletier, commander of 1 Canadian Air Division, said the aircraft’s flight director ended up “in competition with the inputs that the pilot was trying to actually induce in order to set the recovery. That element of conflict resulted because of the flight-control inputs.”

Pelletier said the series of inputs from the pilot created a scenario that had not been tested during development of the sophisticated control system but did not elaborate on what, exactly, the computer disagreed with. Those intricacies will be dealt with in a thorough investigation of the accident but in the meantime, Canada’s Cyclones have been cleared to resume flying. Pilots have been trained in the “very narrow band” of flight conditions they need to be aware of to avoid a repeat of the circumstances that caused the fatal crash. The flight manuals have also been amended. The ship-borne helicopter was returning to HMCS Fredericton after a NATO exercise when the crash occurred. The wreckage was recovered by an underwater drone from water more than 10,000 feet deep in the Ionian Sea and the remains of all the crew members have been identified.

Russ Niles
Russ Niles is Editor-in-Chief of AVweb. He has been a pilot for 30 years and joined AVweb 22 years ago. He and his wife Marni live in southern British Columbia where they also operate a small winery.

Other AVwebflash Articles


  1. The pilot should be PIC. That is the regulation. The software should never, ever over ride a PIC during critical maneuvers that could result in a crash like this. When software is used to over ride commands it must only be used when the flight envelope of the aircraft is exceed by X percentage. I love the new technology in aircraft, it makes the PIC load less and lets us focus on flying the aircraft where and how we want it to, it is our first responsibility. But when conditions demand we can vary from those regulations and envelopes to the extent needed to assure safety for the crew, passengers and aircraft.

    • Except when those excedances would also cause a crash. All we know from this report is that the pilot tried to do something that the FBW system didn’t agree with, but we don’t know what that was or why the FBW system didn’t agree with it. Maybe the pilot got the aircraft in a position where the FBW system determined that the rotor speed would get dangerously low (and when that happens, helicopters become falling bricks), and rather than trying to exit that condition, the pilot continued to fight against the control laws. Or maybe it was something else, but in any case, that’s what the investigation will discover. Until then, it sounds like they have identified the condition that causes whatever this problem is, so it’s something the pilots can avoid.

  2. Crashing for pilot error is bad enough. Crashing for computer error is unacceptable. Fly-by-wire / computer is here to stay and I am sure that it make the pilot workload easier, BUT…..

  3. I don’t think there will ever be a solution for this. Whenever a PIC makes a bone headed decision and people die there’s always a call to put a system in place to ensure that it won’t happen again. And then whenever an automated system causes the same situation there’s a call to ensure that the PIC can always override. So I don’t see how we can ever make it perfect.

  4. I saw the press conference on CBC News World where the RCN/RCAF revealed these preliminary findings. It was a bit difficult to get a definitive picture, because the reporters didn’t know what technical questions to ask, and the military didn’t want to say too much yet. What I got from it is: There was no mechanical failure, but the aircraft did something the crew wasn’t expecting (yeah, it crashed), and the interactions between the pilot’s control manipulations and the flight director/autopilot resulted in a ‘pitch bias’ (nose down attitude?). From my years working on software, I’d guess that the combination of inputs, that were not expected by the programmers, caused the computer to come up with the wrong answer, and it drove the nose down while the helicopter was at a low altitude.

  5. Apparently the “flaw” is in a lack of programming vision by the programmers. That should bring out the lawyers, politicians, and media to demand rules making the programmers liable if they have failed to anticipate every possible piloting, weather, aircraft configuring scenario. Soon, fly by wire certification will include the FAA oversight mandating companies to have the proper amount of programming vision. Of course, that will take decades to determine how the minimum programming vision will be quantified. But it will placate the lawyers, politicians, and media long enough for them to move onto the next news dejour.

  6. Please – in the first place, was there a MAYDAY call about the problem?
    If not, tell me how all this info was obtained – recovered FDR? or what?