A) You might try starting acroread in konsole after changing the encoding of the terminal.
iconv -f UTF-8 -t UTF-16 <(echo 'abc#$%%^')
jschiwal@qosmio:~/547968> iconv -f UTF-8 -t UTF-16 <(echo 'abc#$%%^') >tempfile
jschiwal@qosmio:~/547968> file tempfile
tempfile: Little-endian UTF-16 Unicode character data, with no line terminators
You could echo the password through iconv and then copy & paste it into the dialog.
C) Try both.
I'm not sure which LC_ variable it would be but you can precede a command, setting a variable, as in "LC_LANG=en_US acroread /path/to/file.pdf". Also, the acroread command is a bash wrapper for the real command. There are a number of locale variables used in the script, so I'm sure that acroread uses them. I didn't see anything trying "acroread -help".
If a character doesn't translate from one encoding to another, iconv can't convert the text, however if y ou change the encoding of the terminal, and then echo it, I think if you cut and paste the characters echoed, the byte by byte values may be the same.