Czech accent characters

Czech accent characters

Hi, 
I have an issue with dotConnect for Oracle and Czech accent characters. I do not understand why on one computer I need "Unicode=true" settings in the dotconnect connection string (without it the Czech accent characters do not work) but on another computer Czech accents characters do not require Unicode=true. Regardless of the settings direct=true/false. 
On both computers, when I use Oracle Unmanaged data driver instead of dotConnect for Oracle, the Czech accents works OK.
Configuration both computers: Windows 11, x64, Oracle Client 12.2, dotConnect 10.3.10, NLS_LANG=CZECH_AMERICA.EE8MSWIN1250
Can someone please explain this behavior to me? What exactly decides (in the PC/Windows configuration) whether I need unicode=true in the dotConnect connection string or not? 
Best,
Kamil