DX devices are set up in blocks of 16 or 32 channels. While it is common to assume that because a DX168 only has 8 outputs, the second DX168 would use outputs 9-16 but that is incorrect. A second DX168 unit will use the second “block” of channels which means inputs 17-32 and outputs 17-24. This is true for both inputs and outputs. For example, if you had a DX012 and daisy chained a DX168 to it, the DX168 would still use inputs 17-32 and outputs 17-24 (not inputs 1-16).
It’s actually more intuitive and consistent this way. It doesn’t matter how many I/O channels a particular DX device has (or doesn’t have), it will always account for a block of 16/32 I/O channels. This means the routing of a second box is always the same (17-32) regardless of the type of DX box the second box is connected to. If the system filled in the I/O using the “next available” number, it would be a cluster fuck. Using that method, one day your second box might be using inputs 17-24 and the next day your second box might be using inputs 1-16 based sole on what type of DX box it was plugged into. That method has no consistency and will change with every variation of box used, leading to a cluster fuck.
