Surface sagging

Just to get it out of the way first, loft (normal) is the way to create this surface. When you first encounter something like this your first instinct might be like mine, to go for a two rail sweep but Rhino suffers from ‘surface sag’: if it’s convex going over the struts, it will be concave between the struts. Your next instinct might be to use a three or four rail sweep but no, Rhino doesn’t support this, we must use the network surface command. Yeah right! Go ahead, knock yourself out. Please, please, please let us have a three and four rail sweep. sagging.3dm (43.3 KB)

Hi,

Why not just loft all the profile curves together. Then surface match the 2 surface edges to the profile curves.

Well, as my first sentence says, that’s what I did. The problem with this lofting approach is, it’s just pure luck the number of struts and their position keep the leading and trailing edge (it’s a wing) correct.

Hi Tone - I’m confused- what is wrong with using Loft and then MatchSrf to true up the result to the fore and aft ‘rail’ curves? That looks pretty clean to me. Your middle curve does not intersect all the section curves, but it can be used for a networkSrf (‘sweep3’) and the result is reasonable it seems to me.

-Pascal

Hi,

I tried this with another modeler that has a multi rail command. The results are not very good even after using all the built in options to make it smoother. Using the default settings the results were terrible. As Pascal noticed your middle curve does not intersect the profile curves.

I have yet to see a case that makes a compelling argument for multi rail. In this case pure luck produces the superior result.

Hello Pascal
In this particular case there is nothing ‘wrong’ with lofting and matching. The reason some of the struts don’t intersect with the ‘centre’ curve is because I had to add more struts to get it to work (to get it close enough a match to use match surf). My point is, it’s just guesswork and hope. Whereas if the leading edge, centre line and trailing edge were the rails and the two end struts the shapes, there would be a lot less fk* about. As for the networksrf; the results I get look less than reasonable to me along the trailing edge (though I’m no expert).

Sorry Stratosfear
I almost missed your last reply. All I can say is, “Pass me my gun Curruthers, I’m just stepping into the library; there may be a loud noise”.

Pascal,

One thing wrong with using MatchSrf is it breaks history. Might not matter in this case, but if you are working freeform and want to tweak around for the right aesthetic that’s a pain. If I can slap in a sweep there I can edit the profiles all I want and see the effect on the surface real time. Love to have more history enabled commands in V6 (holy grail–history enabled fillets. Thank you, thank you, for adding the preview within the command, I can fillet lots of stuff I couldn’t have before.)

To return to Tone’s complaint, maybe there is some way to tune up two rail sweep to deal with the sagging. No point in having a three rail sweep if it would still sag, just in more, smaller segments, as it would if you used the three rails to make two Sweep2 surfaces.

I think the sagging is probably due to there being too many section curves - it’s not sagging so much as overshooting the interpolated points, much like an interpCrv with close-together points will.

-Pascal