The impact ball has shown to give excitations in close resemblance with the excitation from a human step. However due to practice and practical measurement reasons, it is interesting to use the tapping machine in low-frequency measurements. Here, the two excitation techniques; the tapping machine and the impact ball, are compared in terms of statistical dispersion. In the AkuLite project light weight apartment buildings were measured using a tapping machine and a (Japanese) impact ball in the low frequency range down to 20 Hz. The results showed that the tapping machine gives more narrow/better confidence interval in the test compared to the test using one excitation point together with the impact ball. The t-test of the consistency of the difference between the impact ball and tapping machine for the same measurement objects shows weak correlation, which implies that the results from the tapping machine are not normally possible to be interchanged with impact ball results and vice versa, without using a correction factor.