Safety Publications/Articles


The multiengine dilemma

Chop and drop, or clean up and go?

You're rolling down a 5,000-foot runway in a Piper Seminole with two people and half-tanks of fuel on board. You rotate, but an engine fails at 75 feet agl. Your airspeed is 83 knots (VXSE), 5 knots below VYSE. Because there is runway left, your landing gear is still down. What should you do?

Two instructors who were in this actual situation crashed on the runway, destroying the airplane. Fortunately, no one was hurt.

It's hard to lay too much blame on the pilots in this accident, because there is no published technique in any official publication for handling this kind of emergency. In the FAA's Airplane Flying Handbook (AFH), it simply says, "If the engine failure occurs prior to selecting the landing gear to the UP position, close both throttles and land on the remaining runway or overrun." Unfortunately, that is exactly what these two pilots did, or at least tried to do.

The problem is that neither the FAA's publication nor the Piper PA-44 manual addresses technique. When those pilots pulled back both throttles, the result was an immediate loss of airspeed and a stall. The proper technique would have been to immediately pitch the nose down, extend the flaps to full, and reduce power as the airplane descended into ground effect.

This flight was a training exercise to see if the Seminole could be stopped in that distance following an engine failure right after takeoff. That 5,000-foot runway length, simulated on an actual 10,000-foot runway, was chosen because a Seminole's takeoff distance over a 50-foot obstacle, combined with a landing distance to land over a 50-foot obstacle, is about 5,000 feet. The CFIs thought that practicing this procedure would give them a better understanding of how it really worked.

Perhaps the pilots were testing that option because they believed a successful climbout in that configuration was unlikely. The Airplane Flying Handbook also says that if the gear is down, "the chances of maintaining directional control while retracting flaps (if extended), landing gear, and feathering the propeller and accelerating are minimal."

That may be true for older, underpowered multiengine airplanes with poor single-engine performance and with an engine-driven hydraulic pump, where a failure of that engine means the only way to raise the landing gear is to allow the engine to windmill for hydraulic pressure or to use the hand pump. It may not be true for modern twins with larger engines and much better performance, as well as fast-retracting electric-hydraulic landing gear systems.

Whether a pilot in a similar situation could successfully retract the gear, feather the failed engine, and continue climbing depends on conditions and pilot skill. An in-flight test of single-engine performance should be conducted during a pilot's aircraft checkout. That test will always tell the truth and should be done as follows:

  • Climb to a simulated hard-deck altitude of 3,000 feet agl.
  • Reduce airspeed and lower the landing gear.
  • Hold altitude and reduce airspeed to VR plus a few knots.
  • Apply full power to simulate a takeoff, pitch the airplane for normal climb attitude, and allow the airplane to accelerate to VY as is the normal climb profile.
  • Upon reaching 50 to 100 feet above the simulated hard deck with an attained airspeed of VY or VSSE if such an airspeed is published in the POH, pull the mixture control to cut-off, failing the critical engine (if there is one) abruptly. (You need to be aware that both the AOPA Air Safety Foundation and the FAA disagree with my recommendation for actually failing an engine, because of the potential for turning a training emergency into an actual emergency.)
  • Adjust pitch to maintain VXSE or VYSE, retract landing gear, and promptly verify and feather the inoperative engine. (Do not use simulated zero thrust as such settings are inaccurate and thus give false performance results, sometimes positive or sometimes negative.)

If the airplane will climb satisfactorily at 3,000 feet agl it will only perform better when at a lower field elevation, given three inches more of manifold pressure plus the ground effect to aid in performance.

The message here is don't always believe magazine articles that say climbout following an engine failure with the gear down is impossible, and that to continue is certain to result in a VMC upset. It just isn't so for a properly trained and competent pilot, flying almost any twin-engine airplane at a low enough density altitude for good performance. If there is no obstruction or rising terrain ahead, just retract the gear, feather the failed engine, climb out to a safe altitude, and return for a normal single-engine landing. You will know by your own tests and the single-engine climb charts whether or not the necessary climb performance is available.

So does the gear position, up or down, make a difference in the decision to continue or abort? Not to me. And it shouldn't make a difference to anyone else if the gear can be moved up or down in seconds. Decisions to continue should be made on known aircraft performance capability, not gear position.

E. Allan Englehardt is a 36-year pilot for a major airline and is currently a senior Boeing 777 international captain. He was Flight Instructor of the Year in 1976, and has been an FAA designated pilot examiner for 15 years, with some 1,200 flight tests given, of which 400 were in multiengine airplanes.

By E. Allan Englehardt

Back to the Index of Instructor Reports