Posting sample data will help in the Community giving you a good answer. One possibility is using the hours, minutes, or seconds functions to convert the data in your timetable to hours (or some units easy to work with), then use the hist function to map the time data into 24-hour long bins. Then each bin having less than your desired number of readings needs to be looked at and used to change those timetable readings to NaN.
See the documentatoin article titled "Dates and Time - Functions" for a full list of functions available for working with time series data in a timetable.
