From the literature I've read ( Al.Rosetti History of Romanian for example ) it looks like we can talk about Vulgar Latin until the 4th or 5th century in the Balkans, and further than that many innovations are specific for an Eastern Romance branch ( Dalmatian as well as Southern Italian languages ). However this changes one the Slavs cross the Danube and occupy the Balkan peninsula in the 7th century, dividing the western and eastern romance speaking Balkan populations. My guess is that this is the point where the ancestor language of Romanian and Aromanian is born. There are however some innovations that take place in Romanian, Dalmatian and some southern italian languages, but not in Aromanian, like "cl" in "clamare" or "clavis" becomes "chema" ( Rou.) and "chiave" ( It. ). Does this mean that this innovation was universal for Eastern Romance but somehow it still isnt present eveywhere in Proto-Romanian? So how is it possible that so many languages developed from Latin in the balkans in such a small timespan: Vulgar Latin > Eastern Romance > Balkans Romance > Proto-Romanian > Romanian and Aromanian ? About which centuries are we supposedly speaking?
In the question the evolution of the Balkan romance languages is treated in the terms of a tree model with intermediate proto-languages at different stages. Some of the perceived difficulties can be resolved by assuming a wave model of language change. Now the archaic features of Aromanian (like the preservation of Latin ⟨cl⟩) are quite natural: It lies at the periphery of the Romance dialect continuum and was not reached by certain innovations that affected the other parts.