do codepages/locales/CCSIDs affect internal representations?

I'm really confused about how/if codepages/locales/CCSIDs affect me.

I'm writing a license key checking routine on the AS/400. The output of
this function is heavily dependent on the actual values of the bytes I
put into it. When I do something like retrieve the system serial number,
are the values of the bytes sitting in the serial number variable at
all affected by the locale the AS/400 owner has chosen?

The AS/400 function needs to agree with a function on a PC that uses
an entirely different code base. If the license key generation consists
of hashing a buffer, do you suppose it will be enough to add one extra
step to the AS/400 function (converting the buffer to ASCII prior to
hashing)? If so, will it always be safe to use iconv to translate from
EBCDIC to ASCII? I know I want to translate *to* ASCII, but I'm unclear
what format to translate *from* if my program is running on an AS/400
with an unusual (to me) locale.