Some thoughts:
(1) Try to have regular meetings with your supervisor - preferably once per week. This will force you to make progress on your project and establish a venue for regular feedback.
(2) Be sure you're clear on the purpose of your project. Why are you constructing this phantom? What similar phantoms already exist? Why can't something that already exists be used for the final purpose of this project?
(3) Establish a clear end-point. What questions do you aim to answer by the conclusion of your project? Does your supervisor agree that these are achievable in the time period you have?
(4) Keep an active list of references. As you progress through your project you will likely come across different references relevant to it. LaTeX allows you to cite references really easily, so learning that and establishing a reference list will help you tremendously. Otherwise, if you're Microsoft-based you might want to look into a program like EndNote (although those aren't free).
(5) Try to start writing stuff up at the beginning - at least your introduction and methods. Even if it all changes by the end, putting in a little thought in the beginning can save a lot of grief later on.
(6) As far as style goes - discuss this with your supervisor. Ideally, I would recommend that you try to write it up as if you plan to submit the work to a journal - so adpot the style that you would see in a journal you would be most likely to submit it too. If you don't know what that is, take a guess and follow that (you can always reformat later). Usually the best hint is to look to the journal that most of your references are coming from.