For loading the endmill, which we already knew would work, we would add a drill stop or shaft collar to the endmill to control the stick out length. But, first we had to remove the endmill, so off we went, rotating the toolholder while heating the end.
Although the heat gun was producing over 450 degrees, the small end of the taper would not rise above 110. For some reason, the heat was transferring to other parts of the assembly. The carbide endmill read over 200 degrees as did the CAT 40 shank. The coefficient of thermal expansion of the carbide is almost half that of the toolholder material, but the hot air gun was heating the carbide more than the toolholder end. In other words the carbide was expanding about the same rate as the toolholder, so it was not going to come out no long how long we heated it.
So, the takeaway was, while we could certainly load the endmill into an empty holder by heating the bore first with the hot air gun and then moving to the outside diameter, we could not use the same technique to remove the endmill.
Sharing information about high performance milling technologies, the result of 30 years of research.