Uncertainties in FFT frequencies

There are several sources of uncertainty in the FFT for getting the peak frequency. Hopefully you all ran multiple trials at each length, so you would have found a slightly different peak fundamental frequency each time, constituting your precision error. The accuracy error basically comes from how noisy the signal is at the peak, and how much confidence you have in picking the 'true' location of the peaks for frequency (this will likely be small). Resolution of the frequency for the FFT is given by a pretty simple formula:

frequency resolution = Fs/N where Fs is the sampling rate and N in the number of data points in the discrete time signal.

Hope this helps.