Preionization is believed to play an important role on the implosion of gas-puff Z pinches. Some experiments have used an external preionization source, e.g., UV light or electron beam. In contrast, other experiments rely completely on over voltage breakdown by the own generator's voltage pulse. However, this approach lacks shot-to-shot reproducibility since self-breakdown is mainly a stochastic process. In this work, we performed a systematic study on self-breakdown using two different cathode geometries: (i) a smooth, round cathode to provide a homogeneous electric field, (ii) a sharp, knife-edge-like geometry to enhance the electric field locally and eventually electron emission. The experiments were carried out on the Llampudken current generator, which provides a current pulse of ∼400kA amplitude and 200ns rise time (10%-90%). We implemented gated XUV imaging, filtered diodes and time-integrated x-ray imaging to obtain information about the implosion as well as the stagnation phase for the two cathode geometries. We found that erosion of the knife-edge cathode might be a serious problem, and we had to replace it every 15 shots. On the other hand, the round cathode lasted for the whole series of experiments. We also measured a more reproducible and larger peak current for the knife cathode. From the photo-conductive detectors we observed that even if the round cathode might present shots with higher x-ray yield compared to the knife cathode, dispersion is almost twice as large. Moreover, after a statistic analysis, it is demonstrated that the dispersion in the yield is due solely to differences imposed by the cathodes and not to variations in the driver, as no correlation was found between them. We found that in order to fit the experimental data with the snowplow model, only ∼60% of the total mass is compressed in the knife cathode while ∼20% for the round one, highlighting the importance of the cathode and preionization. Therefore, we conclude that the use of the knife cathode increases the reproducibility of the experiment in comparison with the round cathode.