ikegami's reply covers the problem as stated in the OP, but whenever you need to convert unicode text to some non-unicode encoding, it's useful to be able to look at it with a general-purpose diagnostic tool, to check for unicode characters that don't exist in the target encoding. Look at this node for one approach: unichist -- count/summarize characters in data (and you might also want to look at tlu -- TransLiterate Unicode).