Normality test for velocity distributions

Hello,

I have a velocity distribution of a number of particles that I want to see if they are normally distributed according to the maxwell distribution in three dimensions. I divided the velocities in each of the x,y and z dimensions and applied some of the known normality tests such as Anderson Darling, Shapiro Wilk, Chi square etc. Now my problem is since their results sometimes differ, how to know which one suits such data the best? Number of particles: 24, and the velocities are represented as an array of 500 elements for each particle. Thanks.