Hello krass,

Aha, that's why the old drawing doesn't make sense, it isn't drawn as built. That said I think the modeling is not as complex as I originally thought. Do you happen to have the specs or drawings for the areas marked in green? My strategy on this would be to build out those surfaces first by 1.5 and use that built out surface as the mating surface after boolean, trim out a sphere appropriately and blend the two parts together. Without trying it I'm not sure how successful offsetting would be, it might just be a manual upper and lower surface process.