This is a long text. Please bear with me. Boiled down, the question is: Is there a workable in-place radix sort algorithm?
It looks like you've solved the problem, but for the record, it appears that one version of a workable in-place radix sort is the "American Flag Sort". It's described here: Engineering Radix Sort. The general idea is to do 2 passes on each character - first count how many of each you have, so you can subdivide the input array into bins. Then go through again, swapping each element into the correct bin. Now recursively sort each bin on the next character position.