Hmm.
Klaus' measurements seem dead right to me, on several armatures and with several ATDs. Always 6mm plus by way of gap.
The tapered section on a K2F (yours has the 18mm bearing at the drive end by the look of it, and if it is a 1954 it should) will be - in round numbers -
*15mm long from narrow end at base of thread to shoulder at the fatter end
*tapr diameter goes from 11mm to 14mm - again in round numbers.
*The threaded section is 11mm long.
*The overall length from face of the brass end cheek to the end of the threaded section is 38.5mm give or take.
(Earlier 15mm armatures are probably exactly the same, but haven't got one to hand to measure.)
If those dimensions are incorrect, then there's a problem with the armature.
I don't think Dave Lindsley has 'done anything' - he'll have worked on what he was presented with and the coil and parts will 99.9% for sure be the ones you sent. He'd have said if not, and there'd have been a discussion about the need for replacement parts.
Had he had the flea-bay ATD you wanted to use with the magneto, he'd probably have been able to point out the problem. If by any chance he did a trial fit of a known-good ATD, I wonder if that threw up any problem? If he did and there wasn't, then the ATD would be the likely area of difficulty. Worth asking the Q maybe.
So, either the armature has a problem with its dimensions or with the state of its taper - I still can't quite figure out why it protrudes only 19mm from the case per your original photo as that is definitely short compared to what I see on mags here - or the centre bolt on the ATD has an enlarged female taper. A small oversize goes a long way on these things as has been said. New bolts are available though.
There will be a solution, there always is, but it needs some careful comparisons with known good bits to get the diagnosis right.
Don't suppose you have a plain pinion or sprocket you could offer up to the armature to see if that goes on and sits about right? I'm thinking that something might have fitted OK, as otherwise it would have been difficult to conduct dynamic tests on the repaired magneto. You can use a couple of nuts to drive the things anti-clockwise, but most people use tapered drive pieces I think.