Summary of pages 1 to 31!
Just read page 1 to 31 - lots of great deductions. I thought i'd summarize what I've read (from an aero engineer perspective)
1. FDR trace matches ATCPS system activating to retard eng #2 to idle + feather, eng#1 bleed off etc.
2. the FDR doesn't seem to indicate any prior reason for the ATCPS to activate (specifically no oil pressure drop or TQ drop).
3. The #2 engine maintains healthy operating parameters (al biet at idle) throughout, and shows a quick acceleration immediately prior to impact (ITT, NH, NP, FF all going up) - nothing to suggest engine has actually failed - at least not completely.
4. There have be previous incidents of erroneous ATR72 ATCPS activation.
initial cause faulty ATCPS?
5. There seems to be conflicting CVR transcripts so we can't really trust any - but at least one translation clearly states the pilots correctly identify eng#2 fail early in the piece.
6. FDR shows PL#1 retarded in stages and shutoff
7. FDR shows PL#2 set to 100% (whoever is moving the levers is moving the wrong ones - that much is clear!)
8. CVR then seems to suggest confusion in cockpit as to why eng#1 is feathered and shut-off - supporting the idea the pilots DID correctly identified end#2 as the failed engine.
9. Crew then attempt restart of #1 (~20s before impact) (again, they think the #1 is the good engine). Eng#1 responds to start but it's too late.
10. PM was used to sitting in the left seat as PF. He was called in to fill in and sat in right seat as PM.
All crew responses make sense IF you swap the levers. i.e. PL#1 with PL#2 and CL#1 with CL#2. Did PM mix up #1 and #2 levers from an unfamiliar seat despite crew otherwise correctly identifying #2 engine/ATCPS failure? (of course it would have been better if PM did nothing at all to either lever until everything calmed down abit)
the hole in this theory is why didn't PF/PM notice unexpected response to PL reduction and correct it? (there is 45s from initial PL reduction until FSO) - my only suggestion here is distraction by stall warning which occurs ~20s after PL reduction and subsequent information overload)
11. In the final moments ATCPS is deactivated (probably since pilots initiate restart procedure for #1 eng, moving PWR management to MCT, which deactivates ATCPS). the #2 eng unfeathers and begins accelerating once again suggesting the engine is actually in not too bad shape - i.e. ATCPS failure, not engine failure. In any case it's too late.
p.s. I suggest speed management and wing drop in the final moments is a bit of a red hearing. The only real chance these guys had in this hostile environment was restore power and fly away, although it maybe fair to argue the initial speed decay leading to stall (warning) was a unnecessary distraction.