Well, this thread has had 774 views, so someone might still be interested to know what happened…
Micro-surgery delivered the two ends of the broken armature winding into a position where they could be slipped into a small-diameter ferrule, with a third wire added, to make contact with the commutator. The wires were bent over as they exited the ferrule and cropped. The ferrule was crimped, the joint was soldered – for electrical integrity – and epoxy was dribbled around and heat-cured, to ensure electrical insulation and in the hope that no movement would take place. The third wire was connected to the commutator, crimped and soldered. That took a couple of patient hours, and a lot of impatient expletives.
Going around the commutator, measuring the resistance at between diametrically-opposed commutator bars showed consistent resistance, except for one position. It turned out to be a loose commutator-winding connection, so it was re-crimped and soldered. Opposite and adjacent resistances all consistent, and no continuity to the armature's body. Yippee!
So the motor was reassembled, correcting some distortion of the flimsy frame caused by use of a bearing puller. Apply mains (with a soft-starter in circuit – because I could). Motor ran up to speed without drama, but when at full speed, the impressive 'ring of fire' around the commutator was present, as before. Bad words. I assume something in the winding moves under centrifugal force, either breaking a widing or shorting to 'earth', or there's an earth leakage I didn't find at low-voltage testing.
New motor required. Pity…
Edited By Kiwi Bloke on 14/06/2020 11:29:54