Hi Sladan,
Is the system an English system?
Configuring the scanner to save FSF files instead of xml.gz will almost certainly solve the problem. You can still get xml.gz files from the XML Enricher - and these should be encoded correctly.
The mechanism used in the UNIX scanners of PDI 7.1.1 for translating data into UTF-8 (when saving to xml) did not always work, resulting in files with invalid characters. This has been fixed in PDI 7.2, which will be available this month.
I hope this helps you,
Allan