Usually 20 minutes are spent on defending a thesis on defense. Decent to keep within 16-17 minutes, in different councils in different ways, so you need to talk your work with the clock in hand. I have performed several hundred times already, and even now, when I need to do some important speech, I utter the pieces of my speech aloud with the clock in my hands. It’s out loud, not about yourself. It helps me to structure the text and understand what I have not finished in my speech. Your speech should be at least 5-7 times.
The text of the speech should be based on the principle: the faster the better. For 17 minutes you never, under any circumstances, no details can’t tell. Even if there are some results that are personally very dear to you, because you have done something, it is not for advice. This is for a scientific seminar, when you have an hour and a half to discuss with your closest colleagues, who understand the subject matter, but it should not be on the defense speech. It is necessary to state the main results, preceded by detailed explanations. It helps me a lot when I give a lecture out loud. It often happens that the speaker tells something and uses concepts that he did not define before. This is a monstrous thing. Everything that you say, everything should be explained, but it is not necessary to touch the general things that everyone knows.
Answers on questions. Very often, when the question is already clear, the dissertator can kill and begin to respond. Do not do this ever, wait until the end of the question when the questioner stops. Sometimes it happens that a member of the council, when asking a question, tells something first 5 minutes. Wait for a calm end, it’s not up to you to interrupt a member of the dissertational council, this is the business of the chairman of the council. Then say: thank you, do not abuse phrases like: what a good question, how I like it, start answering. Under no circumstances should you let the questioner know that he did not understand anything from your report.