First off, if you have audio with your video (that you want to replace with your high-quality audio recording), then step #1 is to nudge the Zoom audio track to line up the impulse of the clap with the corresponding impulse of the scratch audio from the video file. To do that, you need to change the Timecode Display Format to Audio Samples. Once you do that, audio can be nudged a sample at a time, which makes it possible to get two audio files to agree, rather than a frame at a time, which makes it impossible.
To learn the keyboard shortcuts (which include the Nudge commands), print and study this page from the Adobe manual. TL;DR: Nudge right is Alt+Right on Windows and Cmd+Right on OSX. To nudge five frames at a time, add a Shift. To nudge left instead of right, change from the Right Arrow to the Left Arrow.
Once you have your high quality audio aligned with your principal video, it's time to sync all the rest of your video with your principal video. If you shoot at 24fps and capture audio at 48kHz, a single frame of video is 2000 audio samples, which is a lot of samples to sort through looking for the start of the clap! But don't worry about that. Once you have established principal video, all other video should be aligned based on video frames, not audio samples. Unless you shoot with genlock, your video frames are not going to be perfectly in sync, which means that other video tracks could be as much as a half a frame early or late when aligned on a frame boundary. Such is life. Audio that's off by half a frame is not unlike sound that's coming from 21 feet away instead--about a 20ms delay. There's nothing to do about that because (1) video needs to sync to video--you cannot slip it half a frame, and (2) you're not using audio from the other video tracks, you are using your Zoom audio, and that can only be locked to a single reference point.
If you are really OCD about timing, you can average your video offsets, weight them according to which is most prevalent/important, and align the audio so that lies at that weighted average point. I don't know anybody who does this, but if you have two video streams that are 1/2 frame apart (worst case) and you split the difference with the audio, your audio/video error changes from 1/48th of a second worst case to 1/96th of a second worst (and best) case. Only you can tell if that's preferable to you.
