After a great many conflicting news stories, it's finally settled that one recording device has been recovered from the Superjet 100's crash site on Mt. Salak, Indonesia. The National Transportation Safety Committee will download data from the cockpit voice recorder's solid-state memory boards this weekend.
Here's a photo of the CVR, which searchers found among debris from the tail empennage:
It looks like the FA2100 model from L3 Aviation Recorders. I feel confident in guessing that the data is intact. Even though the jet flew full speed directly into a cliff, because the flight data recorder and cockpit voice recorder are in the tail, that permits a good deal of deceleration prior to impact.
It doesn't matter that in general the unit looks pretty beat up. The only part that matters is this large dome-shaped unit, snipped from the official photo:
That's the crash-survivable memory unit, or CSMU. It's got a stack of flash-memory boards inside. It's probably intact because it's designed to tolerate pressurized seawater (down to 20,000 feet), sustained fuel fires, 3,400 G's of impact, and a wide range of chemicals. The dome provides three layers of protection: aluminum on the innermost surface, an inch of ceramic thermal insulation, and a quarter-inch of stainless steel on the outside. Given the small diameter to be protected, that adds up to a lot of protection.
The damage to the metal housing to the left of the CSMU ...
... is irrelevant, because it holds circuit boards for transferring the four channels of voice data into the CSMU. The boards aren't needed after the aircraft is destroyed. Same for the base:
... which serves as a structural chassis, and a convenient location for data and power inputs. It also holds a circuit board.
What about the little tube-like item on the right?
That's the underwater locator beacon, or ULB. It's a pinger that starts sending out acoustic pulses at a frequency of 37.5 kHz upon immersion in water. Here's what a ULB looks like in happier times:
The ULB is the focus of some interest in the event-recorder community, because (as I mentioned in this post) neither ULB attached to Flight 447's data and voice recorders was detected by the pinger-search teams, even though their ships passed almost directly overhead and during the nominal transmit duration (30 days).
I asked a flight data recorder expert about this minor mystery and I gather from his comments that while the cause isn't known, recent ULB tests at a lake in France show that the metal frame of a CVR can "shadow" the transmissions, drastically reduce the detectable range in those directions. Added to the great depth of the Flight 447 wreckage (more than 12,000 feet), it could explain how the pingers could transmit but not be heard.
It's too early to say what happened in the cockpit prior to the crash, but there's been plenty of speculation in the interim: that the captain and first officer were distracted by promotional duties, perhaps even admitting VIP's onto the flight deck to gaze out the windscreen. If true, that conversation will be preserved on the audio channels.