About being a good (physics) teacher
If you were able to start over as an undergraduate again, what kind of professor you would choose to study/work with? Or, if you were to become a professor, what kind of professor you would like to be? This kind of questions recurred to me from time to time. So whenever possible, I am always interested to observe how different professors teach their courses and lead their groups.
I believe that a good educator has to be well-trained and well-prepared. I probably got that belief from my earlier unsuccessful mentoring experience. More than 10 years ago, started from my college years, I always wanted to give my younger sister (then in high school) and other younger relatives some useful suggestions in terms of study and career plan etc. However, since I myself didn’t have much experience so I had little to offer, although I really wished to be a good mentor. At that time, my sister was thankful for my encourage and supports, but probably also felt bored by my useless suggestions 
Back to the topic of professor/undergraduate student issue, here are my opinions about being a good (physics) teacher.
About teaching: When teaching a course, always give a detailed syllabus. Tell the students the reason that you choose the particular textbook; what is your strength in teaching this course, e.g., the parts related to your research; what do you want them to grasp from this course and the connections with other parts of the field. If you know there are some good online sources, like lectures videos by some wonderful teachers, ask the students to watch and read. Sometimes, a study group is more useful than class teaching in terms of answering specific questions of a student.
About research: Encourage collaboration between students while also reward individual creativity. Should let students know, that any advance of science is based on a lot of work by others, and if they are interested in doing science they should learn to collaborate with each other and learn to initiate collaboration (as in other businesses). When doing real research project, for most undergraduates, the first project is better to be a small topic, and should be in collaboration with senior members (gs, postdoc, prof) in the group. Publishing a first-author paper certainly can be a strong incentive for an undergraduate student, but without input from others, the process can be very slow and painful. For an undergraduate with limited time, he/she can collaborate with other students and finish one project together, and all as co-authors. They should be informed that the main purpose for an undergraduate project is to get a taste of doing research and understand the physics. Usually only for simple projects exploring new fields (e.g., nano stuff), can they make substantial contributions. For more difficult projects in those more matured fields, the undergraduate’s involvement probably won’t contribute too much, or even may delay the progress.
One additional thought about the collaboration with others: besides participating student organizations, I think the ability could be also acquired from participating sports like soccer etc. I myself like to play soccer, and feel organizing an offense attack is pretty much like collaboration in a project, although for me most of time the former is easier and more fun