Traditional DAQ is limited to returning SGL or I16 binary data. A Single-Precision floating point number is 32 bit, but only 22 of the 32 bits are actually used for representing the numeric information, which means that it can represent data from a 447x board more accurately than a I16 number, but it still won't represent the full 24 bit accuracy of the board.
I would recommend programming the boards in DAQmx because it can stream data to disk more efficiently, and it supports using 32 bit integers for binary data. There is a shipping example in LabVIEW 7.x which demonstrates streaming binary data to file with DAQmx called Cont Acq&Graph Voltage-To File(Binary).vi. This example can be found in the example finder under Hardware Input and Output>>DAQmx>>
Analog Measurements>>Voltage.
Good Luck,
Logan Kunitz