On a slightly different tack, I thought it might be useful if I shared a recent experience when I was reconditioning a couple of 3 rail Hornby Dublo locomotives. I've cobbled together a test bed (as illustrated) using some meters which I had intended to use on the control panel of my analogue "scale" model railway, but decided not to use. The rig consists of a 0-20 DC voltmeter and 2 ammeters; 0-1amp and 0-5amps. Unfortunately they are not "centre off" so I needed DPDT switches to give readings in both forward or reverse. The track is 4 feet of Peco Code 100 with a third rail added via copperclad sleepers. The slide switch at the end simply helps feed power to either left and right rails (for 2 rail locomotives) or to both outside rails and the centre rail (for 3 rail engines).
The first subject was a 2-6-4T with a very weak magnet which barely ran at all. Not having access to a remagnetiser I fitted a neo magnet and the engine ran well but was pulling about 1 amp at 5 or 6 volts. I put a multimeter across each pair of commutator segments and this gave equal measures of resistance across each of the armature windings - which suggested that the armature wasn't the problem. I loosened the gear wheel and tested that the chassis was free running. No problem. To cut a long story short, after cleaning the commutator with isopropyl alcohol, ensuring correct gear mesh, oiling everything that needed it, and many test runs,all of which showed 1 amp current consumption, I cleaned out the slots in the commutator with a fine toothpick. Immediate transformation. Power consumption now settled to between 0.4 and 0.55 amps depending on load.
The second subject was an 0-6-2T which I picked up as an indifferent runner. The gear wheel was bone dry and there were traces of rust on the armature shaft. The armature shaft was also tight in the bearings. I got it running satisfactorily but, like the 2-6-4T, it also pulled nearly 1 amp. This time I went earlier to the commutator. The slots were very close and needed a fine brass wire carefully manipulated to clean them out, followed by a cotton bud and IPA. Again, the amps dropped to well within the 3-6 range.
It seems to me that, had I not tested the current draw, and then picked up on the carbon build up in the commutator, I might have been happy with the running of both locomotives. And both may, over time, have cooked their armatures.
By way of a final comment, these Dublo mechanisms are a pleasure to work on. Everything is repairable or replaceable and, as been pointed out before, will be running long after some of the more "delicate flowers" from today's manufacturers will be in the 4mm equivalent of Woodham's scrap yard.