In this paper the breakdown time delay td in neon was measured and the formative time delay tf determined in three different ways: (1) from the Laue diagrams, by taking the values where linear approximation of experimental data intersects the time axis; (2) from histograms, by taking the minimum values of delay times $t_{d\min}$ for the formative time; and (3) from a difference $t_f= \overline {t_d}- \overline {t_s}\approx \overline {t_d} - \sigma\,(t_d)$, where standard deviation $\sigma\,(t_d)$ is approximately equal to the mean of the statistical time delay $\overline {t_s}$. The dependency of the formative time delay on working voltages Uw was found and fitted by different approximate analytical models. The separation of the formative and statistical time delay distributions was discussed and the distribution of the formative time delay experimentally obtained. The nonstationary exponential distributions with time dependent parameters were introduced to describe the statistical time delay distributions under conditions of increased cathode sputtering.