Easily reproduced here for pretty much any "ae * 0+-3" so clearly something is screwy with rays coming in perpendicular to the tgc vector.
Attached image shows ae 300 1 on white where you can see the root solver failures just starting on the top-grazing edge and around the circular end-cap on the right.
Easily reproduced here for pretty much any "ae * 0+-3" so clearly something is screwy with rays coming in perpendicular to the tgc vector.
Attached image shows ae 300 1 on white where you can see the root solver failures just starting on the top-grazing edge and around the circular end-cap on the right.