The most definitive approach would be to conduct a test.
To a measured volume of naphtha, add a carefully measured amount of DMT freebase such that it will all dissolve at room temperature. Place this in the freezer and after the appropriate passage of time, collect, dry and weigh the precipitated freebase.
You now know how much freebase remains in the naphtha post freeze precipitation and can use that figure to calculate how much acid you'd need to use in order to recover it, how much base to add, and the minimum amount of fresh naphtha you'd need to pull it from the basic solution using three pulls.
By evaporating that naphtha you'll find out how efficient (and/or worthwhile) all this extra effort actually was.
You could just cut to the chase and evaporate some (i.e. a measured volume) of the used naphtha that you already have and use the amount of DMT obtained as a guide to the amounts of acid, base and fresh naphtha to be used in the recovery process which is essentially the same as the the test process outlined above.
At some point you have to either evaporate the naphtha (or another solvent...) or decide what to do with a little bit of DMT salt because, obviously, repeated freeze precipitations of ever decreasing size invokes the law of diminishing returns fairly early on.
As far as the arithmetic of how much DMT will react with how much acid, etc. I would urge you to look through the forum as this is something I've explained approximately 97 times during my time here at the Nexus (I know, I should have just programmed it as a macro button - I'm not that organised [world's worst chemist]).