To characterize the pulse cleaning of textile and rigid filter media in technical filter houses and test rigs, by now mainly tank pressure and valve opening time have been used as characteristic parameters. The pressure pulse formed, however, depends on the whole geometry of the plant so that these operation parameters do not allow a comparison between different technical systems. Thus, the question is raised-also for the development of new pulse jet systems-which parameters of a cleaning pulse are decisive for its performance. By knowing these parameters, an experimental setup could be adjusted in a way that it shows the same cleaning behavior like a filter house so that the laboratory filter tests are comparable to technical filter cycles. On a filter test rig which can be used for both textile and rigid filter media as well as on a pilot plant designed for Herding alpha filter candles, experiments concerning the influence of these different parameters on the cleaning behavior were performed. By inserting different orifices into the blow tube and varying the tank pressure and valve opening time, cleaning pulses of great variability were generated and the history of pressure drop over the filter medium or with candles even on the filter surface was recorded. The performance parameters maximum pressure pressure integral and pressure rise velocity of each pressure pulse were determined and compared with the cleaning efficiency achieved. As result of comprehensive regression analyses we found that especially the maximum pressure and the pressure rise velocity during the zero passage of a cleaning pulse are decisive for the cleaning result. The practical conclusion is that pulse jet systems should be optimized with regard to these parameters. The pressure integral however is not a significant performance parameter.