Engine Mix-up Led To 737 Ditching


The NTSB says the crew of a first-generation Boeing 737 that ditched in the ocean off Honolulu on July 2, 2021, mixed up which engine was underperforming on the plane and steadily reduced power to the good engine while trying to coax more out of the damaged one. The aircraft went down a few miles offshore and both pilots survived. According to the final report, the Transair cargo flight had just taken off from Honolulu with the first officer flying when they heard a thud and the pilot flying correctly assessed that the right engine had lost some power. As the crew worked the problem, the first officer reduced power on both engines to slow the plane to a target speed of 220 knots and subsequently mixed up the engines and told the captain it was the left engine that was affected. “The captain accepted the first officer’s assessment and did not take action to verify the information,” the report said. 

The captain took control but remained under the incorrect assumption that the good engine was on the right. He added throttle to the damaged engine and it responded somewhat but not enough to keep the plane in the air. Meanwhile, the undamaged left engine was near idle and he did not adjust the left throttle. There wasn’t enough power to keep the plane in the air and he ditched in the ocean. The report says the failure to verify which engine was affected was “likely the result of the captain’s high workload and stress.”

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. If you are unable to tolerate the level of workload and stress required to verify an engine problem (basic multi-engine stuff) in a 2-crew environment, perhaps commercial flying is not for you.

  2. I think this will trigger some additional scenarios in the simulator, like “FO identifies wrong engine”. It’s frequent that trivial changes could have prevented a crash. In this case, push both throttles forward until you figure stuff out.

  3. Retrained some guys who survived a similar misidentification in a combat zone. In their case they correctly performed the procedure’s memory items but then failed to run the entire checklist, eventually shutting down the healthy engine. They became distracted by the sensor operators’ loss of composure, couldn’t clearly see engine indications because of smoke, and took remedial measures applicable to past aircraft but not the one they were in. Bottom line, things can go very wrong in an emergency, especially is the crew does not use the tools provided in the checklist and fails to verify each others’ actions and perceptions. In my case and this one, they were very lucky to survive.

  4. Similar situations have happened before, a DC-6 carrying a load of Christmas trees out of MIA in the 60’s (feather #1, mixture #2, firewall shutoff #3), crashed on 36th Street. Then Air Florida 737 into a bridge in DCA while operating at reduced thrust because of iced up PT2 probes on two perfectly good engines. Boeing solved part of this problem on the B-777 by silencing engine fire warnings on takeoff until reaching 1000 ft agl. In most emergencies, you either die right away, or have plenty of time (the most likely scenario), so a good rule is in case of an emergency is to sit on your hands and count to ten.

  5. Wow! There is a lot in this report and I haven’t even finished reading all of it. But let me pick on one topic – communication.

    It seems that the rest of the World’s airlines find it useful to ‘cut through the language barriers’ by saying “Mayday, Mayday, Mayday” while the U.S. system still uses “declaring an emergency.” This seems to be problematic for U.S. pilots when flying overseas with foreign controllers, for whom English is often a second language, but evidently even the HNL controller couldn’t quite grasp the problem when this Captain ‘declared an emergency.’ Nor did the controller seem to grasp the meaning or significance of the word “standby.” It isn’t the first time time either, with well-intentioned controllers interrupting flight crew as they are performing safety-critical tasks, as can be seen in this video: https://www.youtube.com/watch?v=4UHWl62szvY

    News for controllers: flight crews also get busy, especially on an engine failure right after takeoff and we don’t need your help until we tell you we need assistance or are ready for your help. The “Mayday” call in this case is not a literal00 translation of the French “m’aidez” (help me) but rather is notification that we have a problem, we are invoking the Captain’s emergency authority, and we may be (and probably will be) deviating from a clearance to ensure safety. The best you can do is clear the path for us.

  6. Had both crew members done dozens of engine out procedures in the simulator they might have done the right thing.

  7. Meanwhile, a crew on a NormalAir 737 was able to identify the real engine and landed without incident.

  8. When I was an A&P with PHI in Lafayette, Louisiana in 1990 at the refurbishment hangar there was an incident whereby one of our pilots ditched a float equipped Bell 212 near an oil rig. He lost an engine and instead of going to 104% power on the good engine proceeded to shut it off and had to put it into the Gulf. When the ship came to us on a flatbed truck he’d crushed, like an eggshell, the belly skin which was subsequently replaced. Oh, yeah, and the Chief Pilot fired him.

  9. So one engine has a problem and you are still losing altitude while playing with the throttle. Push them both up then figure it out!

    • Yes. In simulation we frequently don’t do that because of fear of damaging someone else airplane engine. I took a CFI ME checkride in an A65 Queen Air years ago. During the simulated engine failure go around, we just kept settling. Was still lots of MP available above the MP green arch limit on the good engine. But couldn’t use it because of fearing of overboosting the aircraft owner’s good engine. That training mindset may carry over into a real engine failure situation. .

      • Well over-boosting might be better for the engine than ramming tons of sea water down its throat when you ditch

  10. Why in te world when they realized they were in trouble didn’t they just push BOTH thrust levers forward??? And the “target speeed of 220 it’s?? What is that?? Drift down speed?? Drift downs are accomplished with the operating engine at MCT…what in the world were these two doing???

  11. In my experience when training single engine ops in a multi engine airplane, some turboprops and jets are much easier to deal with than a piston twin. Other than verifying the failed engine (or auto feather) there is nothing much more to do than to fly the airplane to a safe altitude. The one thing I noticed on the report is that there was no verification of failed engine securing. In other words when reducing the failed engine thrust lever to idle the monitoring pilot audibly verifies each action to secure the failed engine before moving anything. This would avoid shutting down the good engine and associated systems. No need to be in a hurry, unless unable to extinguish a fire. Not familiar with Boeing procedures or systems, but in the sim the jets I have flown a single engine out even a V1 cut is an easy event to deal with compared to piston procedures. The Hawker 800 I fly is even easier due to the rudder bias system it has.