I recently discovered a couple of neat tricks using the DT25 and Helix. If you connect the Helix using L6 Link, the preamps are disabled on Channel A on the DT, but they are available on Channel B. Using DTedit via MIDI, you can disable the preamps on Channel B but you can still use the cab modeling. This allows you to toggle between Channel A with no cab modeling (essentially just the natural DT speaker with no added cab modeling) and Channel B with the DT cab modeling enabled. You can actually assign 4 different cab models to the 4 'voicing' slots on Channel B, but you have to manually switch between them via the DT front panel, as there's no access via Helix/L6 Link.
Similarly, you can use the internal DT reverb on Channel B, and assign different reverbs to each of the 4 slots, even with the preamps disabled. This allows for greater DSP flexibility on the Helix in case you run into that particular limitation.
So after that elaborate setup, here's the question:
Where in the DT signal chain does the cab modeling live? I'm assuming it's after the power amp modeling, but that's not entirely clear.