If you actually conceived the new invention yourself (i.e., you didn't copy it), and you were the first to do so, then basically all you have to do is provide a written description of the invention that would enable a hypothetical "person of ordinary skill" in the relevant art(s) to make and use the invention.
(There are other requirements too, of course - for example, you have to disclose the "best mode," as then subjectively contemplated by you, for practicing the invention; you have to provide at least one written claim, which in essence is an infringement checklist; you have to disclose to the patent examiner all "information material to patentability" that you know about; you can't wait more than one year after the first offer for sale, public use, publication, etc.)
Here's the underlying assumption: The first inventor should be allowed to set up a toll booth for 20 years (originally 17 years) as a reward for "merely" documenting how to practice the invention in a centralized national database, that is, in the Patent Office collection.
That might have made sense 200 years ago. But I've long wondered whether that part of the patent law has outlived its usefulness -- arguably we don't need the same incentives to help build up a national library of engineering knowledge as we did in 1790.