Thanks for explaining your views on how TNG-tech would be applied to the older ball-turret phasers; after thinking it through some more, I do think this is the best explanation for what we've seen.
The notion of phaser banks vs. phaser emitters is a problematic one. Here's a crazy idea: what if there's something special about phaser power? What if the energy that powers the phasers is drawn off the main power grid, and then converted and stored in the ship's actual phaser banks, which are basically phaser "batteries" (in the electrical sense.) The energy has to be pre-phased or something to charge the phaser banks, but can be de-phased and re-routed back into the main power systems in a pinch. Then, the energy is phased again as it is passed through the emitters. As crazy as it sounds, this might solve a few of the problems we've seen. It might even explain those phaser power cells from DS9.
Perhaps the difference with the "fed off the warp drive" phasers from TMP was that energy was converted on-the-fly for phaser use, possibly in an attempt to not "waste" energy sitting in storage batteries. It makes me wonder if all ball-and-turret phasers worked this way, or if it was just a dubious flirtation on the refit, not carried over to other ships.
Slight tangent, what do you make of the "megaphasers" on the Miranda
? I had contemplated at one point making the things on the Excelsior
's interhull the same thing, especially considering how close they are to the intermix chamber.