Based on this information in your "Anonymous" reply above:

I am using the so called MSSQL Management Studio Express to actually see the data in the tables. I can paste 'real' Unciode data into it through the interface, as I can run an INSERT-statement with the N'unicodedata' syntax, and it works fine.

The next question would be "what encoding is being used for unicode characters in that tool?" The point is that the difference between UTF-16 and UTF-8 is critical -- converting character data back and forth between the two is a fairly simple and common task, but if you make the mistake of treating one type as if it were the other, everything goes to hell.

I haven't used MS tools enough to figure out which ones use UTF-16 vs. UTF-8, or how (or whether) they provide flexibility in this regard, but I do know that MS tools can be prone to use the encoding that you might not expect.