I've done some comparisons for time taken to manifold render between calculations using my rounding over method compared to Minkowski. The results are listed below,  rounding first, Minkowski  second   calculate(s=5,knob=60,fil=6,thick=15);              3.8secs     8.3secs   calculate(s=5,knob=60,fil=10,thick=20);            6.0secs        9.6secs   calculate(s=12,knob=60,fil=6,thick=15);            8.5secs    11.2secs   calculate(s=27,knob=100,fil=11,thick=24);       27secs 29secs The rounding 'steps' are 0,1mm, which do not show when sliced/printed. Higher resolution is easy to set, but it will take longer to render. Another advantage, is that the rounding profile can be adjusted, to allow a steeper start for the bottom edge, eliminating supports, for example. However, I think Minkowski may be a better option, if it is intended to scale the stl file, or produce it by something other than by fdm printing. On 06/01/2025 01:03, Raymond West via Discuss wrote: > There is a formula, but a bit tricky, as I said, a number of triangles > involved. The 2d line drawing I posted a few days ago gave an > indication of the construction of the basic shape. I completed it for > the simple case, where the scallop diameter equals the diameter of the > 'blobs' lying between them. > > There are only four parameters required for the basic shape, for > example - calculate(s=5,knob=48,fil=6,thick=15)  (s is number of > scallops, knob is overall diameter, fil is radius of fillet (top and > bottom), thick is total thickness.) It works fine now, from one > scallop up to whatever. (I argued with myself wrt one scallop, whether > or not it should produce  no knob, but I settled on a simple circular > one - if the scallop size was smaller, then there would be no argument > 🙂). > > The more scallops, or the larger the knob, the slower the Minkowski. I > modified my bevel script, to do round-overs, I think that would be > quicker, than Minkowski, but would depend on the radius of the > round-overs. (The above values for 'calculate' takes less than ten > seconds to calculate using manifold, and 25 seconds for 25 scallops, > both using Minkowski.) > >  If there are different radii for the scallops compared to the > 'blobs', then the triangles between the centres are no longer > equilateral, but Isosceles, which makes the calculations a bit > different, but not much more so,  I've not bothered with that, > although it would be the generic formula, it would introduce a couple > more more parameters, unnecessary choices when none need to made if > all we are concerned with are generic scalloped knobs for fdm printing. > > On 05/01/2025 21:45, Thomas Richter via Discuss wrote: >> Hi all, >> >> I wish you all a good start into the new year. >> >> I took a few days off and worked on it only sporadically but today I >> finished the analytical solution (almost - see problem description >> below). The base body can be created in the current state of the >> software. The object in the images attached here renders in about 500 >> milliseconds. >> >> It can be found here: >> https://github.com/trses/screw-nut-knobs/blob/webapp-proto/analytical-solution.scad >> >> @Raymond: I am convinced that there is not one (rather simple >> trigonometric) formula that can generate the shape, but two for the >> two different arcs. You also have to determine the points of contact >> of the two arcs. It can of course probably be solved with a >> Taylor-like approximation but I didn't want to go that way. >> >> This afternoon I realised that I'm almost done building a general >> solution to make any polyhedron out of an arbitrary list of layers of >> points. They points of one layer can even have different z values. I >> might go those few final steps sometimes but at next I will build the >> web application to generate the knobs. >> >> If anyone has the nerve to look at this, I'd be interested in >> suggestions on how the features of OpenSCAD could be better utilised. >> I have a very software engineering heavy approach to tackling these >> things. >> >> Additionally, I realized a strange problem: when setting the number >> of arms (variable ARMS) to 13 or 17 the following error occurs: >> >> WARNING: PolySet -> Manifold conversion failed: NotManifold >> Trying to repair and reconstruct mesh.. >> ERROR: [manifold] Input mesh is not closed! >> ERROR: [manifold] Surface_mesh -> Manifold conversion failed: >> NotManifold >> >> There are a couple more value combinations where this happens but >> there seems to be no general rule. For larger numbers of arms like - >> e. g. - 20 it works again. If you comment out the subtraction of the >> hollow sphere in line 147 it renders without problems (but looks >> strange though ;-)). I have checked the polyhedron several times by >> hand, it is generated correctly. >> >> OpenSCAD seems to have a glitch with boolean operations on more >> sophisticated shapes, possibly especially when there seem to be very >> small resulting faces. >> >> It would be nice if someone had an idea about this. I could easily >> extend the body and generate additional layers for the surface to >> omit the subtraction at all but I'd like to understand what's the cause. >> >> Thanks again for your support. >> >>> Am 26.12.2024 um 20:06 schrieb Adrian Mariano via Discuss >>> <discuss@lists.openscad.org>: >>> >>> Ray, are you saying it's only the upside-down rounded edges that you >>> think are "awful"?  I don't really understand.  I basically round >>> every edge I possibly can in my models---but that doesn't include >>> upside-down ones.   I'll have to look into that adaptive layer >>> thickness. >>> >>> On Thu, Dec 26, 2024 at 2:00 PM Raymond West via Discuss >>> <discuss@lists.openscad.org> wrote: >>> >>> On 26/12/2024 15:03, Adrian Mariano via Discuss wrote: >>>> I'm not sure what Ray has against rounded edges.  I use them all >>>> the time and think they look and feel great in my FDM printed models. >>>> >>>> What does NOT look great in my experience is something like the >>>> gradual dome top of the knob, which tends to have a very exposed >>>> layer line step-structure with very wide and visible steps that I >>>> think look and feel pretty bad.  That will be worse the closer you >>>> are to flat---I don't know how bad it would look with the model as >>>> shown.  I printed a knob to go on my audio amp and my first version >>>> had a domed top like shown and I scratched that and went back to >>>> flat after I felt the model in hand. >>>> >>>> The other thing that does NOT look great is a roundover printed >>>> upside down. >>>> >>> Hi Adrian, you half answered your own question to me, and nop head >>> gave  the top half answer,  wrt adaptive. >>> _______________________________________________ >>> OpenSCAD mailing list >>> To unsubscribe send an email to discuss-leave@lists.openscad.org >>> _______________________________________________ >>> OpenSCAD mailing list >>> To unsubscribe send an email to discuss-leave@lists.openscad.org >> >> >> >> _______________________________________________ >> OpenSCAD mailing list >> To unsubscribe send an email to discuss-leave@lists.openscad.org > _______________________________________________ > OpenSCAD mailing list > To unsubscribe send an email to discuss-leave@lists.openscad.org