LDCad 1.4b (win+linux)

[Philippe Hurbain]

I recently generated a lot of RJ cables with LDCad and noticed that twisted cables didn't always looked nice. So I made a little experiment to see what happened.
   
From top to bottom:

• Folded cable, no twist. The random color shows the nice refinement of segment length
  
• Straight cable - perfect of course
  
• 180° twisted cable. Looks completely degenerated. Probably some singularity in calculations
  
• 180° twisted cable with one end slightly raised. Unaligned end avoids singularity but twist is concentrated in ends segment instead of beeing evenly spread.
  
• Adding a 90° rotated control point in the middle gives a much better look, but there is no refinement of segment size.
  
• Same thing, but middle CP is offset to cause refinement. Looks even better...
  
• When limiting the twist to 165°, there is no longer singularity but the low number of segments makes it look rather coarse.
  

I remember Kevin Clague telling me that he had a hard time with torsion when he wrote LSynth...