Progress has been...promising so far.
I used a brand new box cutter blade and tried carving up some pink antistatic foam. I keep this stuff around for storing and transporting the very large boards my Silicon Graphics machines use and I find chemically it's quite stable. A lot of experimentation found that after cleaning out the old foam and slipping in the new foam with tweezers I could reduce the commutator gap by about 1-1.5mm. Not a lot but given how close the brushes were previously when the foam wasn't expanding this should be enough to achieve a proper contact. Initial tests on the bench at least seem so show that there is a resistance from one brush lead to the other.
I really wish I knew what the original foam pieces looked like or how rigid they were.
Edited: Further searching has found this problem is NOT exclusive just to the 2400.
It happens as well on at least one other model of tape machine, the AV-3400 as well. (read 3/4 of the way down)