I managed to isolate the thing in my main model that makes the difference and causes the tree rings. It uses $fn on the main barrel because I had found (with the 2015 stable version) that tuning $fa and $fs was not working beyond a point but $fn seemed to get around that problem. It uses the same $fn value for the rotate_extrude of the end dome. So attached is a test model that produces tree rings with no DXF involvement. If you set BARRELFACETS to zero then you get a smooth end. The issue is not quite the same as the one with my DXF test. With the DXF, the tree rings happen without $fn on the rotate_extrude and are of unequal size. In the latest test, they need the involvement of $fn and seem to be of equal size. Too late for my print though – the STL went for print yesterday, so sanding down will be happening. Regards, Andy