Air France (F-GITA) Accident Tahiti (1993)

[cf6-80c2b5f]

This is an old accident, but I am researching to discover at what point the software logic was changed on the B744 with regard to the sine qua non of this accident.

At the time of the accident, if the autothrottles were left engaged as you crossed over the MAP (E/D) on a VNAV approach, the FMA logic would shift to THR REF||VNAV SPD and the power levers would surge forward to do an "automatic go-around" of sorts. The Air France FO manually overrode the throttles, but after touchdown, when he went to reverse, the throttles started forward again and he missed the #1 reverser, allowing the #1 engine to be at go-around thrust while the other three were in reverse. They ended up in a lagoon off the right side of the runway.

Somewhere along the line, Boeing must have changed this logic because the Boeing Flight Crew Training Manual now states that on a VNAV approach, "VNAV path guidance transitions to level flight once the missed approach fix is passed." FCT 747 p.5.48 (June 30, 2015).

I can find nothing in any manual that discusses this automatic go-around feature. I suspect Boeing realized this was a bad idea, so at some point after the accident they revised the software. I just don't know when. Can anyone point me to any documentation (memos, bulletins, etc.) that would shed more light on when the changes were made? Thanks.

[tdracer]

CF6, I wasn't involved in this investigation, but I knew a guy that was. IIRC the problem was the pilot pushed TOGA then changed his mind and continued the landing. The pilot claimed the engines didn't respond to the thrust levers, but as you noted he'd missed one and they went off the end. (but again it's been 23 years, my memory may be faulty).

[cf6-80c2b5f]

Thanks for the reply. The FO didn't press TOGA -- the aircraft automatically went into a quasi-go-around mode after passing the MAP. From the accident report:

It is part of the design of the automatic flight system that if the End of Descent point is reached with the system still active, the automatic flight system concludes that the visual approach has not occurred and, therefore, the missed approach procedure must be applied. The characteristics of the VNAV (VOR) mode are then the computer-based equivalent of the FIR approach procedure.

When the automatic flight system initiates an automatic go-around, the autothrottle commands increased engine thrust to obtain the reference thrust displayed by the EICAS while the flight director gives nose up instructions to maintain the speed displayed on the MCP (or the calculated safe speed). ...
In the case of this accident, the pilot did not follow the flight director’s instructions. As he used manual controls to reestablish a path with a descent rate of 500 feet per minute and pulled the controls back towards the idle position and held them there, the increased thrust command was maintained, as the reference thrust had not been reached and the autothrottle was not deactivated.

[JammedStab]

The only thing I can think of is, if you don't get any info in this thread, try to find someone who knows how to get you a copy of an old FCTM. Perhaps a retired 744 guy.

[cf6-80c2b5f]

Thanks, Jammed Stab. The oldest I have is 2006 and there is no trace of this automatic go-around feature in it. Apparently, even when it was in effect, hardly anyone knew about it. In typical Boeing fashion, they can't seem to write a manual that is understandable by the average pilot, as the accident report reflects:

a) The pilots were not aware of this feature of the VNAV mode.

It is worthwhile noting the following points:

• Boeing and Air France pilots qualified on the B 747-400 who participated in the technical investigation were not aware of this feature either. It was necessary to call on the expertise of specialists in the Boeing design office to obtain this information.
• Boeing’s documentation did not mention this feature (some information has since been added), nor did Air France or UTA’s.
• The captain indicated that when he went through qualification training on the B747-400 at Boeing, instructors told him that the VNAV mode operating in VOR mode still had some glitches to be ironed out. He had since received no additional training on the subject. We can thus see that the particular point in question was neither studied, nor introduced, in the course of other exercises.
• Air France instructors were not aware of this feature of the airplane and were therefore unable to include it in their training of the copilot or in periodic training of pilots.

Here's the link to the report if anyone is interested. http://lessonslearned.faa.gov/AirFrance747/accident_report.pdf

[bigal cessna]

As a lame licensed on early boeings but not on the b747 -- i was told the reason the qantas boeing 747 speared off the end of the runway in bangkok ,,,, was that 0n the early b747 you could reverse three engines with one engine lever forward of flight idle stop-- ---- but of the latter b747 all throttles had to be on flight idle stop , to remove the reversing interlock , thus allowing the reverse levers to be raised --- this could be the reason when he went manual ------- as for boeing changing the aircraft systems , at qantas on the first b707 , one training captain trick was to apply the parking brake lever on approach and carry out a normal landing ---- but he came a cropper when the next UPDATED boeing -338 arrived , same trick ,, but 8 main landing tyres burst , AND RIMS GROUND DOWN ----- ouch TOOK SOME TIME TO FIND NEW WHEELS, AND TO GET THEM TO THE TRAINING AIRFIELD IN THE BUSH , TO GET IT IN THE AIR AGAIN --