Hi, I started designing guitars to go down the cnc route. Just started to evaluate Rhino (win)and a few other programs before I buy. I quite like the Rhino interface so before I go much further I thought I’d try some curved tops. I’ve done a few models that I found easy like tele and strats etc, and even necks on anything I find rather straight forward on Rhino. When I create a guitar top thats flat its no issue, just tried using fillet edge and surface blend, all gets good results, BUT I’m having a problem with joining curved panels to create a seamless join when viewed in render. I’ve tried to keep the curves simple, rebuilt them, etc etc and using sweep 2 is getting the best results. Heres a few screen shots to explain and a simple practice model that I made to evaluate if I could get this to work. I do know that this will CNC with no issues but purely as a modelling evaluation for me, I’d like to see if I can get this join perfect.
generally you can use MatchSrf which will let you work that continuity. but since both surfaces have a bit of a difference in control points, (probably the input curves did not have a matching number of contol points before sweeping), matching may not result in perfect solutions. and you have chopped all the curves into segments with a g1 continuity at best, at the back the unsurfaced part you even have a g0. maybe you have the initial curves before splitting, then sweep the entire guitar in one go, or use network surface and simply trim the last part with the red curve off. this red curve actually looks pretty messy, so you have to rebuild if you need this one for the final shape.
Hi Tony - it looks like you have drawn the perimeter curve for the shape in several segments corresponding to the lateral curves - that is ok but you need to make sure you have good (G2 at least) continuity between segments - yours appear to be tangent - for this type of swoopy shape, I’d keep things curvature continous.
Also, I would try making this ins a radial arrangement and not in lateral sections. And with greatly simplified curves:
Yeah MatchSrf doesn’t do a good job, as you say the control point are close together. The curves that created the two surfaces in the picture both have 12 control points, in fact all the vertical curves have 12, the smaller joining ones 6. I read a few comments on here and they mentioned that so I did that and tbh it made it better. Regarding the curves is there anything I can look at to learn how to make better curves? So splitting ruins the curve, is that what you are saying? How would I sweep that design in one go, sorry for the lack of understanding on my part…
I know for sure there’s something in my method that’s wrong here…
Hi Pascal, you know I can’t remember but I think It was one then I split it tried a few surfaces then used blendcrv to join it all, here’s where I might have joined them tangent… doh, so tell me could I sweep the entire shape using sweep1, is that what you mean?
Hello - you might be able to Sweep2 - the example I made was 4 separate EdgeSrfs from simple curves - the point count is manageable if you want to make adjustments to the surfaces, which you will - but mostly I am suggesting a a layout: the center panel which I am guessing is flat or very nearly flat, and the body surfaces that bridge between that and the outer edges - those swing around the roundy end of the center panel - this way the curvature of the surfaces is distributed sort of radially around that center panel - I think you can keep better control of the shape that way, especially if you keep the surfaces simple - yours look way more dense than they need to be and that makes life difficult at the seams between them as you found - matching surfaces only deal with three rows of points at most, so if those are all crowded near the edge, the match may be made ‘technically’ but still look like hell.
Hi Pascal, I’ve done another model which has the flat portion on top then sweeps around solve the curves perfectly, for example a PRS style or Ibanez, however this model has no flat portion at all. Its hard to get a method that suits this one. I’ve only ever seen a couple of this style of model done correctly and I’ve seen loads, most are not very good. I can only seam to sweep this model in postions which is why I broke it up. I stand to be corrected of course but one sweep I dont think can be done here. hmmm not sure what to do…
BUT not a requirement with our subd… we are mathamatically totally ok with ngons and triangles when used with purpose.
And by totally ok, I mean the model will not break like tsplines used to. It may not give you the exact shape you were looking for, BUT there are circumstances where a well placed triangle or ngon will work just fine (such as a fading bevel detail) . It will also convert to nurbs no problem.
