Adding MLLP-characters before encoding content causes the MLLP characters to be messed up in at least UTF8. Latin1 and ASCII works well. https://github.com/dib0/NHapiTools/blob/master/src/Base/Net/SimpleMLLPClient.cs#L113