Robert Moore wrote:

James Robinson wrote

The drop in airspeed was unnoticed, and the stall seemed
to catch them completely by surprise.

I wonder what the stall warning was doing all of this time?

The FDR data is posted on the NTSB web site, in both graphical and text
form. Here's a link to the graphical data:

http://www.ntsb.gov/Dockets/Aviation...027/417236.pdf

Looking at the graphs, the aircraft settled in at 2400 ft pressure alt
(1650 ft radio alt) with AP on in heading mode. After leveling off, 5
deg flaps were selected, and the throttles were increased to about 30%
torque. Over the next minute, the airspeed slowly climbed from about 160
kts to about 190 kts.

The next significant event was the throttles were dropped to about 10
percent torque, and 5 seconds later gear down was selected. At the same
time, the AP switched from heading mode to LOC mode, as the localizer was
captured, still maintaining altitude.

Airspeed was initially dropping at 2 kts per second, and increased to
about 3 kts per second as the gear extended. The airspeed dropped to
about 130 knots over the next 25 seconds, and at the same time, the AP
steadily increased pitch from about 3 degrees to 10 degrees to maintain
altitude.

The next events happened in rapid succession:

- Ice detect alarm on (message flashes on display).
- Flaps increased to 10 degrees.
- 2 seconds after flap select, at about 130 kts, the stick shaker
activates (It is set to activate on low airspeed, indicating impending
stall.)
- 20 lbs. pull is shown on both left and right control columns in
response, and throttles are increased to 75% torque.
- Aircraft pitches up sharply, reaching 30 degrees over the next five
seconds.
- As the aircraft pitches up, the AP shuts off.
- 2 seconds after the start of the pitch-up, the aircraft rolls sharply
left, and the stick pusher is activated. The aircraft rolls 50 degrees
over 2 seconds. The stick pusher is triggered by high AOA, and indicates
that the aircraft has stalled.
- Both the rudder pedals and control wheels are in a neutral position
while this is happening, so the sudden roll is likely because of
differential ice accretion and the left wing stalling, as a guess.
- In response to the roll, the wheel is moved to the right, and the right
rudder pedal is pressed. The aircraft rolls from left 50 degrees to
right 100 degrees in 5 seconds.
- While the pitching and rolling is going on, airspeed is sitting at
about 100 kts, and the FO retracts the flaps on her own initiative.
- The right roll is overcorrected, and the aircraft rolls to 40 degrees
left, and then again rolls right to 100 degrees for a second time.
- Finally, the roll is stabilizing at 30 degrees right, but the aircraft
has pitched down by 45 degrees. They attempt to pull out of the dive,
pulling up to 2Gs with 170 lbs combined pull on both control columns.
(120 on left, 50 on right) but run out of altitude.
- Airspeed at the end of the recording is about 130 knots, pitch down at
25 degrees, power still at 75%, gear being retracted. Altitude loss is
900 feet in the last five seconds, with no sign of abatement. Overall
time from start of stick shaker to end of recording is about 25 seconds.

So where does this all end up? The crew was inattentive to the speed
loss, and everything hit them at once. They reacted the wrong way to the
impending stall by pulling the nose up and applying power, ending up with
a true power-on stall. Retracting the flaps and gear was also
questionable. The captain overreacted to the rolls making things worse,
and unrecoverable.

The NTSB is probably going to come down hard on the airline on crew
selection, training, and management oversight. They will also hit the
FAA for not having sufficient training and check requirements.

Overall, this will probably be a watershed accident for the regional
carriers, where many rules regarding hiring, training and proficiency
testing will be substantially tightened.