some observations on drift analysis and ULBs
With the two recorders retrieved and at least one of the two Dukane ULB's, it may be interesting to revisit some of the material on the earlier search phases.
It is striking that the "retro drift points" of the Brazilian Navy and of the US Coast Guard as mentioned in BEA's December 2009 interim report (page 80) are both less than 10NM away from the actual impact zone. They used the same model (NCOM - Navy Coastal Ocean Model, page 79). With hindsight quite accurate, taking into account the complaints about absence of sufficient reliable data on winds and currents close to the accident date.
The 2 June 2009 BEA Interim report states (page 46) that the ULB's had a limited signal propagation range of "two kilometres at most". Based on GPS data, the METRON analysis (figure 17, page 20) establishes that the US Navy TPL, tugged by the Fairmount Glacier, went 8 times over the 10NM (~ 18,5 km) quadrant (J30) of the impact zone before 1 July 2009 (i.e. within the 30 days of the ULB's certified operating life) with a "spacing of 2,5 km" (page 48 of BEA's interim report of June 2009), to take into account the "scan swath of the TPL which is approximately 2 NM" (= 3.7 km). This in turn is in line with the METRON report, which states on page 21: "The TPL sensors were assessed to detect the ULBs at a lateral range of 1730m with a POD of 0.90" (2x1730~3,45 km). On page 47 of the BEA interim report of June 2009 it is stated that the TPLs had an "average detection range" of "at least" two kilometres per TPL trawl. Presuming that the TPL, as foreseen, was trawled close to the (flat) ocean floor, this operation would have had a higher than 0.90 probability of detection at the time. Certainly now it seems that both recorders have been lying around unobstructed on that same (flat) and relatively firm ocean floor.
That is of course if one or both ULB's had functioned as per their certification. The METRON report also states that based on a series of aviation accident that ended up in water, it is relatively rare that one or both ULB's do not function. The estimate of a 90% survival rate for the ULBs may itself be low for a crash at sea that does not involve a fire, the report states.
We all know that the TPL operation was unsuccessful. It seems therefore of importance that the ULB's are inspected for possible reasons of failure. One has been retrieved with the CVR. Finding the one belonging to the DFDR may still be a priority for BEA.
PS @ grity: the photo you posted with the DFDR and CVR mounted on top of each other does not reflect the situation on board of AF447. The METRON analysis clearly states (figure 16 on page 19) that the CVR was mounted between FR71 and FR72, while the DFDR was located between frames 83 and 84.