Here is the recorded RDAS data for the flight.
I post-processed the recorded RDAS flight data using a Kalman filter. (See my NARAM 44 R&D report.) This filter combines the pressure and acceleration measurements along with a model of the rockets motion to produce an optimal (in a least squares sense) estimate of the rockets state.
The first plot shows the flight from just before liftoff till after apogee.
The next plot zooms in on the time around apogee..
Notice that the velocity estimate passed through 0ft/sec at 19.27 seconds. This is a much better estimate of apogee than either the RDAS or AltAcc integrating acceleration algorithms provided. The AltAcc fired its apogee charge at ~24 seconds.
The primary reason for both altimeters being so late were the vehicle dynamics. During the motor burn the rocket oscillated rather wildly in pitch. So much so that it effected the pressure data between 4 and 5 seconds into the flight. Because of this pitch oscillation, there was a substantial angle between the rockets direction of motion and the direction it was pointing. This meant that the motor thrust was not acting entirely along the axis of motion which resulted in over estimating the rockets peak velocity.