A disruption of a tokamak discharge is a sudden loss of confinement, or thermal quench, in turn resulting in a quench of the plasma current. The fast release of thermal and magnetic energy could result in very large thermal and electromagnetic loads on the surrounding structures, such plasma facing components or the vessel, especially in large devices such as JET and ITER. Understandably, considerable research efforts are dedicated to develop both timely detectors of these events and mitigating actions. Magneto-hydrodynamic (MHD) instabilities are often seen as precursors to disruptions. The growth of large, overlapping, magnetic islands is thought to be behind the destruction of the flux surface structure that provides the plasma confinement, triggering the thermal quench. The detection of these modes is used to predict disruptions. Usually the analysis of these instabilities focuses on how early and at what level they can first be detected. This paper will investigate a different but related question; is there a specific maximum perturbation level that triggers a thermal quench? This study provides experimental insight in the processes that may trigger tokamak disruptions. The perturbation amplitudes that trigger thermal quenches in JET and ASDEX Upgrade are compared and the results form a strong physics basis to determine protection thresholds to be used at future devices, such as ITER.