On the F-M Erie-builts, I have a couple of articles in Train Shed Cyclopedia #64 for reference. One was a reprint from Railway Mechanical Engineer (RME) for 1946 November, and described the UP 3-unit, 6000 hp combination. This had belt-driven auxiliaries, from which I’d assume that the main generator was the GT567A1. From RME 1948 May was an article on the PRR 3-unit 6000 hp combinations. These had gear-driven auxiliaries, which points to the GT567B1 main generator. They also had Commonwealth cast inside equalized trucks, whereas the UP combination had fabricated trucks with outside overhung equalizers.
Sweetland, in the Diesel Era Train Master book, shows a diagram for a GE-equipped T-M with static excitation, which had belt-driven auxiliaries. Unfortunately there is not more detail, and no corresponding diagram for the Amplidyne version. But given that the GT567C1 was a modified GT567A1, with no mention of a change from belt-driven auxiliaries (unlike the GT567B1 case), then it certainly looks as if it was developed for the T-M, and used initially for the Amplidyne version. It could then have been used for the static excitation version without further internal modification, and changes to the auxiliary machines mounted upon it, but perhaps not the mounting details of the casting.
I wonder if the apparent reversion to belt drive was because the gear drive was not wholly satisfactory when used with the F-M engine. When AEI first used that style of auxiliary drive with the Sulzer 6-cylinder engine, in the Nigerian 1400 class, there were gear failures due to torsional vibrations from the engine. I think that the solution in that case was the use of resilient gears, but I am not sure. The F-M sequence of belt-drive, gear-drive then belt-drive would be consistent with gear-drive being tried and found wanting. But then the Westinghouse equipment first used on the T-M did, I think, have belt-driven auxiliaries per that manufacturer’s normal practice, so that could also have been a reason for its choice when the change was made to GE equipment.
More generally, belt-drives did seem to have an ongoing place in locomotive practice, so they were evidently reasonably acceptable, at least for smaller units. GE used belt-driven auxiliaries in its small export Universals. Alco used belt-drive sin its DL531, and again in the DL535, even though the latter had the GT581 main generator, more usually associated with gear-driven auxiliaries.
Notwithstanding all of that, we’re no closer to the ascertaining the exact nature of the GE GT567, and how it compared with the GT566, also designed initially for 2000 hp input applications. Given that the GT567 appears to have had an upper rotational speed limit of 900 rev/min, and that that limit was set by the maximum armature peripheral speed (I think that 10 000 ft/min was a high-side number back in those days), then it might well have had an armature of larger diameter than the GT566.