You don't understand what encoding means and works.
As You know, data in our devices is written in binary code, but to make it a lot easier readable, we converting them to hex code. So, in UTF-8, ASCII symbols using one byte, while other two bytes (e.g Polish symbols taking two). In Windows-1250, always only one. PS2 homebrew supporting only pure ASCII e.g from Windows-1250 and similar. Another thing is, that OSDSYS using his own codepage (so FMCB also) which is ASCII + markers for another which exist in its font. If You mess with this, You get gibberish in menu or even FMCB reject parsing file (depend of mismatches between what You did and what can be used). As You can see, the same symbol can be written differently in binary.
Another factor is how enter symbols is written. There are three common types: 0A, 0D and 0D0A. CNF parser support only Windows type, which is 0D0A. Everything else used and You lose enters which leads to unexpected by me results.
And now, in some of Notepad version in Windows 10 it starts support UTF8. And it is possible that it writing it by default since some build (as Win10 and 11 are rolling releases...) or You not intentionally switch to it by some keyboard shortcut. I guessing because I don't using it.
The best for You is stop using such limited applications like Notepad and switch to something more advanced like Notepad++, EditPad Pro, SlimlimeText etc. etc.