Many sources talk about the field inensity (and therefore arcing/breakdown) on the outside of the dome as a limiting factor on the highest voltage that can be possibly reached on a VDG machine. There are also references to possible dielectric breakdown along the belt itself. I am wondering if there could be another phenomenon at play. As the belt runs downwards after delivering its charge to the upper roller + brush, it has to pass the inner edge of the window at the bottom of the globe. However carefully we smooth out this edge, there will be a strong field between the aperture and the belt. As the voltage builds up on the shell, there will be an increasing tendency to spray some charge back onto the downward running part of the belt, which will then carry the charge away as fast as the upward motion brings it in. (This process would be stable and continous at some point, and is not the same as breakdown of the belt material). Is this mechanism really a practical factor to consider in VDG design? If so, are there cases where this process would kick in well before the breakdown on the outer surface?