You asked us for this as an enhancement request a few days ago?!:
http://forum.hauptwerk.com/viewtopic.php?f=5&t=6162
See also this similar enhancement request:
http://forum.hauptwerk.com/viewtopic.php?f=5&t=6112
Note that you can also potentially get Hauptwerk v3 to swap the left-right panning for ranks assigned to a group of two mono audio outputs by selecting the 'Cyclic within octave, octaves and ranks cycled' channel allocation algorithm for the group.
However, be aware that, in that case, which ranks would end up being assigned C/C# and which C#/C depends on the RankIDs in the sample set. See also this excerpt from the v3.20 release notice:
BUG HW-000977: Multi-channel audio algorithm updated to give better channel spread regardless of organ definition contents (Advanced Edition only).
The 'Cyclic within octave, octaves and ranks cycled' multi-channel audio channel allocation algorithm (on the 'General settings | Audio output groups' screen) previously determined the starting audio channel for a rank based on its rank ID specified within an organ definition file in order to ensure that existing ranks were not re-allocated to different outputs if a rank was added or removed in an organ definition (for example using the Custom Organ Design Module). It was assumed that ranks would be numbered sequentially by sample set producers, but not explicitly stated as a requirement. However, not all existing sample sets do number ranks sequentially, which could lead to ranks not being distributed amongst audio channels as efficiently as possible. Because various existing sample sets also do not number ranks in constant increments, there is no reliable way to determine a general pattern from the rank ID. Hence the algorithm has been updated simply to cycle the starting channel using an increment of one for each subsequent rank in rank ID order (but not as a function of the rank ID value directly). Please note that if you use this multi-channel algorithm for a multi-channel audio system then you might find that some pipes are now routed to different speakers within their groups, and there is the possibility that you may wish to re-voice them as a result (if you had adjusted voicing according to the previous speaker routing scheme).