I work on mainframe where (almost) everything is EBCDIC*. I often FTP a dataset (file) from mainframe to PC – like a report – so I can email it. I always specify TEXT. My (our) FTP translates from EBCDIC to ACSII just fine.
* even mainframe is *slowly* moving from EBCDIC. Headed to UNICODE. DB2 is leading the way. The current version of DB2 on z/OS, the SYSIBM tables are all UNICODE. So maybe, eventually, we can dump all this translation and just live in a UNICODE world.