I would also say that if you're working on building a "new" encryption system without the knowledge that can be gained from working with existing ones, you're likely to make a lot of mistakes that have been made before.

I would suggest instead of trying to build one of your own at this point, pick up one of the books Avis187 suggested (I highly rate Applied Cryptography for a good introduction.) Try to make your own implementation of some of the more simple algorithms in there and understand why they do what they do. This will help you better understand what to look for in speed, security and complexity. They're all trade-offs.

Good luck with it.