Jay,
Each individual fid is added sequentially to the ser file by the pulse sequence, and what the ordering represents is defined by the pulse sequence. There are two issues here: The detection mode in the indirect dimensions, and the ordering of the indirect dimensions for 3D and higher.
The detection mode for older data (pre-XwinNMR3) is defined by the processing parameter MC2, which will be found in the proc# file. This is an enumeration which will tell you whether the data are Magnitude, States, States-TPPI, echo-anti-echo, etc. In newer data MC2 has been replace by the acquisition parameter FnMODE, which allows the same pulse sequence to be used for different detection modes. It is similar to a combination the Varian "f1coef" parameter and the "phase" array. SpinWorks (and TopSpin) are set up to look for FnMODE first, and then use MC2 only if FnMODE is not set.
The ordering of the indirect dimensions is defined by the AQSEQ acquisition parameter, which can have the value 312 or 321 for 3D data. It is normally set automatically by the loop structure of the pulse sequence. Older data may use the AQORDER processing parameter. TopSpin is set up to look for AQSEQ first, and then use AQORDER only if AQSEQ is not defined.
This is all well described in Bruker's XwinNMR and TopSpin manuals. Also, each pulse sequence should contain documentation on the detection mode and the ordering of the dimensions. I am just starting to add 3D processing to SpinWorks, so I may be able to give you further guidance in the future.
