Glad to hear you got it solved! I did some analysis of this before you got this resolved, so here it is in case this is helpful:
The problem originated from the spline curve on the sketch. Near one end of this line, there seems to be some inconsistency with the points & tangency handles, causing the curve to have a slight overlap:
This was propagating into the revolved body, which was what caused problems for shell.
I was able to resolve this by redrawing the spline, simply clicking the existing spline points from the previous curve (attached f3d file). This avoided the overlap issue, and the body shelled successfully.
However, with this new spline, the shell only went as far as ~2.1mm. Beyond this thickness, the offset encounters another issue, which is a region of the offset becoming so sharp that two curves become 'separated' (the below diagram shows the offset just before it becomes separated):
This occurs because this region becomes sharper and sharper (i.e. the angle between the arc and straight line becomes smaller and smaller) as you offset by larger distances. You can improve the shell distance here by reducing the sharpness of this angle on the input body, or (even better) removing the sharp edge completely: i.e. if you add a small fillet here, you get a smooth offset, which can be shelled much further:
(If you don’t want to have a fillet here on the outside of the model model, one option is to do the above, then delete the fillet after shelling the model.)
Hopefully this helps in shelling to the required distance.
Something else (not directly related to the shelling failure): is it intentional to have so many points along the long sketch spline? The potential problem with this is that this introduces waviness into the surface (this would become apparent when manufacturing):
Perhaps this was intentional? If not, in general, it’s best to use as few points as possible to define a spline curve - this allows the spline to flow more naturally, giving smoother results. Below shows a redrawn spline using only 4 points, giving a much smoother resultant revolved surface:
Hope this is helpful; but if not, feel free to ignore this