I would incident the modeling source. You might have to use some type of ND in the halogen if the ratio is too large, or increase modeling lamp to adjust ratio. Can you adjust snout so that light source wont be directly at the camera? Can you have it light from underneath? Do you have a spot meter to measure glass lit by snout that will let you exclude light source from reading reflected? I have a seconic 508. I would incident the modeling light, ( turn off modeling light ) then take reflected spot reading on the illuminated star, then the light source for the star. That way I know what the ratios are and how far out the star light source will be a wash.