Detrimental effect of a small amount of ripple in a metal halide system

A question of how much ripple should be allowed to be passed from the ballast to the lamp at acoustic resonance frequencies is currently under debate in the lighting community, especially in the standards committees. We have developed and presented a metric for measuring ripple and other frequency content generated by an HID ballast in a way that takes into account the relevant physics of the lamp as an improvement over the prevailing metric of measuring current or voltage ripple magnitude. In this paper, we make an application extension to this metric by measuring the maximum temperature of the arc tube instead of detecting visual changes in the arc appearance or electrode attachment. A bad case scenario is concocted and compared with a near miss to underscore problems with single case studies being used for stability confirmation testing. We chose experimental parameters using our knowledge of perturbing acoustic modes. We have experimentally chosen a sensitive arc tube. Thermal imaging is chosen as a sensitive non-biased perturbation determination. The bad case injected ripple frequency was chosen roughly from peak identification and refined with experimental tuning. Less than 1% ripple at 39.2 kHz on top of a square wave is shown to negatively affect this metal halide lamp. This work is put in context with past experimental and theoretical efforts.