Conceptually, to pass the Turing test you should be able to fool any human (including expert judges). It's just practically easier to choose an expert judge as opposed to test on every human on Earth, since you'd hope that if an expert can't do it, then neither can anyone else.

Also, it shouldn't matter if the human is someone who worked on the AI, or has read the code, or has seen every previous Turing test that the AI underwent. There shouldn't be any information a person could know that would allow them to tell that it's an AI.