You nailed it jaz -- The work had already been done adding the choke, low voltage DC supply, new SS rectifiers, etc., etc., and it simply would have meant tearing all of that out and rebuilding it again on a new chassis, for a lot more work.
From the standpoint of the umbilical, using that approach would have saved only one lone conductor in it to achieve the same operation as achieved with the existing approach, so either approach is basically a wash from that standpoint. So what would that approach have produced?
Solving hum problems means following the well established process of ensuring good lead dress, shielding, ground systems, low noise tubes, power supply filtering, and even potentially modifying the design to have the small signal tube heaters operate from a DC source -- and that is exactly the process that was followed here as well. Invariably, the problem is completely solved at some point in that mix or if not, then some point was not adequately addressed in that mix.
Ultimately, that's exactly what happened here, where the inadequacy was in shielding. But we always think of that in terms of cables, lead dress, or maybe missing covers. We'd never think of it in terms of the final effort required here, as such large scale noise control measures surely would have been taken care of at the design level, right? Except that in this case, that's EXACTLY where the biggest compromises were made. There was simply no place to run to get away from the contamination produced by the compromised power transformer. The only option left then was if the circuits could not be moved far enough away from the transformer, then move the transformer away from the circuits. Since all of the other noise reduction efforts had already been done, shielding from the power transformer -- as opposed to from the power supply proper -- was all that was required to finish the job.
In this case then, removing the entire power supply would have gained virtually nothing, other than to produce more work, and conformity to a typical "understanding" of the way such things are done. On the down side however, remoting the entire power supply would have made for a much larger power supply unit, than can be had by simply remoting just the power transformer itself.
So the answer then was shielding the power transformer, rather than the power supply. Isn't this exactly what other manufacturers did in providing complete end bell and Faraday Shielding transformer construction, and mounting the transformer on top of rather than through the chassis, which was often made of much lower magnetically active material? I've effectively accomplished the same thing here, by using space as the shielding tool. Since the transformer itself was the only noise source left, removing just that component then was just as effective as if the entire supply had been removed as well.
Dave