I have created a simple arched profile shape (A) using InterpCrv in ortho view, gumballed it upright, copied it forwards, B C D, scaled its height using scale 1d.
InterpCrv quad its tops, straight line the base, all travelling A to D.
carried out sweep2 and the last two profiles C and D do a crossover sweep !
Now as I had a real life sweep (also attached…yet to be CurvatureGraphed by the way) which has done such a crossover on every profile, I thought I would experiment and look at display of Dir to see if direction was to blame.
Wish to know what is classic cause of intertwining sweep ? like dogs in and out of poles at a police training yard !
Instead the test that should have gone perfect also twists !!!
It hasnt helped me understand causes as it should have been fine so is there something else wrong ?
TwistedProfiles_Job.3dm (388.6 KB)
Simple sweep2 with clones shape twists on CD.3dm (154.7 KB)
The first two notes in the help file on this command say:
- Make the degree and structure of the rail curves match each other.
- Place cross-section curves so the endpoints are on the ends of the rails or on edit points of the rails.
Neither of these are adhered to in these files. (Note that in my testing of both files, the first point of these does not necessarily seem to matter). You have a rail curve that is nowhere near your cross-section curves.
By the way, Loft will give you a nicer surface on these curves (though the point spacing in the TwistedProfiles_Job file will mess this up rather badly anyway).
Hi Steve, on the first example I found if you pick the straight line first, using sweep2, the result has a twist but picking the top curve first gives a good result - don’t know why.
The other example, as Wim says, you will get a nice surface if you loft - but you should rebuild the curves first using as few control points as you can, these input curves are just too messy for a good result - and generally pick curves on one side to avoid the twist.