Originally Posted by
bearfoil
RR NDB
Yes, easy to say, but it brings up a flaw in the Search strategy. At the outset, and as early as possible, the BEA trumpeted, "Intact at Impact". Then the reliance was on ACARS to entertain "Continued Flight" into weather, and all manner of Radar excuses. This is how it must be read. A "Search" must include ALL areas that are possible, even those that fly in the face of the "Reputation" of what are arguably collateral players, pilots, union, Thales, AB, etc.
[...]
Keeping 10nm away from LKP is (was) inexcusable, boneheaded, and incompetent.
Where are you getting this from ?? Who kept
away from LKP and when ?
From what I've read, the first two search phases were all covering zones "centered on LKP". Phase 1 mostly looked for pinger signals, plus (as I read it) some additional Phase 2 ROV searches to cover areas where they'd run out of time on Phase 1. In addition, they did complete bathymetry of the area to inform future searches - 40NM circle, where ? - centered on LKP, again.
Only in Phase 3 did they do something different -
only after having failed to find anything with a general search centered on LKP, they searched an area calculated from drift analysis of the wreckage. That turned out to be wrong - but with so many variables and uncertainties and the delay in finding the floating wreckage, that's not really suprising. But surely it was worth trying?
Phase 4 went back to square 1 effectively, looking for wreckage rather than pingers, and starting from... - LKP, again.
So in what way were they ignoring or avoiding LKP or assuming the plane flew on for longer ? They centered everything on LKP!
Why again is the DFDR inside the fuselage and not the Vertical Stabiliser ??
Don't they (VS) have a habit of falling off in flight ?
More seriously, it's because some time ago someone decided that tail was most survivable place (probably is in most accidents - not this one though), and they need to be somewhere accessible for maintenance I believe. Someone also decided there should only be one recorder (of each type), very well protected. Probably right in the days of bulky and expensive foil and tape, maybe not now, when gigabytes of solid state memory is a few $ and smaller than a fingernail.
We could probably all do a better job of designing a recorder system now, with hindsight and infinite retrofit budget, but lets not forget that the current CVR/FDR system actually has a very good recovery rate when you consider the number of possible ways to crash a plane.
[ ULB failure in this case definitely needs looking at - doesn't look to be any difficult terrain reason for failure as was suggested before ]