When pre-machining a gear the teeth are being created with a hobbing process in which the hob and gear blank both rotate at corresponding rotational speeds. The hob which has a V shaped tooth profile cuts through the gear blank which is also rotating at a certain speed. The combined rotations of these two generate two profiles. The first being the involute tooth flank (which can be created quite easily using law-curve). The second being the generated tooth root with a variable radius.
For CAE purposes it is important that this generated tooth root geometry is exactly the same as created during the real hobbing process.
My question is what could be the best method for modelling this tooth root profile in NX.
See below for an examples of the generated tooth root profile and the generation hobbing process.
years ago, I had a .lps program that ran in Autocad. Called HOb Mastr Now called Gearzilla but this is old also. They guy who developed this retired back in 2009. I can send you the program I have for Autocad if you want it. At that point you could import this into NX and extrude it.
It may not be pretty either, if I remember it would make many many small lines. Which when you try to extrude the shape it would have many little tiny surfaces/
Perhaps you could use fit curve -> spline through the line end points to create a smooth curve through the result.
@WGE The math for generating trochoidal fillet is know (see appendix B of the attached pdf). When I was developing the code for spur gears I had planned to include trochoidal fillets but found that I did not have access to the hob tooth fillet radius. I assumed that this fillet radius is standard for each given gear modulus but that it varies across the different modulus values so I stopped at using fillet radii.