.. It is relevant at this point to note BDF p57 section 107, on the interchange of ἄν and ἐάν: "Ἐάν appears very frequently instead of ἄν after relatives in the NT, as in the LXX and pap
ἐὰν even occurs after relatives in some Attic authors, mostly in Xenophon, but they have been expunged by editors. cf. Jannaris 421, who blames Demosthenes and Xenephon for the cause of the conflation. See also Conybeare and Stock, Grammar of the Septuagint, §105. They give numerous examples. But no one seems to say why?
Perhaps ἐὰν was often pronounced as ἄν? See also Moulton-Turner, vol. 2, §38 p. 92. , vol 3, §2, p. 108, , Moulton suggested that the result of an accent change from pitch to stress and the slowing down of pronunciation in Koine allowed hiatus of dissimilar vowels to occur (but this is not the case with ἐὰν). This was different than Attic times when smoothness and flow was more important. Moulton says the reason for the mixup is not phonological.
The structure often occurs followed by the subjunctive, and Blass Debrunner cites that scribes may have gotten confused (an orthographic error) or that because of the 'conditionality' of the clause, ἐάν was written and began taking over.
Perhaps Carl can chime in for the reason.