The characters you see are somewhat irrelevant and will differ based on the text encoding used by any given program. Most programs for US Windows are going to use the Windows 1252 code page, which provides a paritcular mapping of glyphs (visible representations of characters) to binary values (bytes, or sequences of bytes if the encoding is multi-byte or variable byte.)

So if you spit out a sequence of bytes in any particular program, you might see different results according to whether they are using Windows 1252, Unicode, ASCI, Latin 1, etc... code pages. All that really matters is what the binary values are.

If you are trying to send some binary sequence from a small controller, and you can only issue ASCII characters and it's really only ASCII characters, then you may be screwed because ASCII doesn't define any characters with the high bit on. There are variations on ASCII that do and your controller might use such a variation.

Most such systems should provide some sort of 'escapement mechanism' by which you can generate untypable characters, such as the %xx scheme used in one of the examples above, where XX is the hex value you want to send or \xYY, where YY is the hex value you want to send. If it doesn't, you may be out o' luck.