You need to configure charset_table or indexing won’t work but it’s quite hard to do.
In my app, search matches should broadly the same as MySQL utf8mb4_unicode_ci, e.g. ‘motor*’ should match ‘Motörhead’.
I wrote Collation-to-Charset-Table convert any MySQL collation to a charset_table. But it does a lot more, including customizing the included/excluded code-points and tuning the folding rules using a human-editable file.
I hope it can be useful.