for these 2 libraries,
I tested their Rijndael DCB/CBC algorithm with \"AES Kn
Are you sure it fails? The vectors are written as hex strings. Did you feed the hex strings as strings or did you convert them first as binary data. If you fed them as strings it is no wonder you get a different result.
NB: I do not use both libraries but my own components but they bioth have a good track record and failing these simple tests seems highly unlikely.
I know another lib, but not if passes your test or not:
DCPCrypt http://www.cityinthesky.co.uk/cryptography.html
TurboPower LockBox 3.1.0 has been released. It includes DUnit tests which validates the correctness of its implememtation using the official AES test vectors.