@@AndrewJacksonDesignStudio I still feel a perfect g2 curve transition between the line and arc is the best. Anyway, after lozenges maybe try g3 ball corner cubes! :D
Pretty sure this has smoother curvature transition than a g2 transition. Also have to balance between eroding too much of the arc and having the transition being too abrupt.
@@AndrewJacksonDesignStudio continuing my point above, this perfect g2 “transition” curve I’m talking about also works great when you have to make offsets. Offsetting any line and arcs with g3 curve between them leads to discontinuities which when you try to correct will result in variations in offset distance between the entities
Really great video as always, cheers A J! I go about this using a similar method in PTC Creo/Style (ISDX), but I don't use the section of arc for the top profile driver. Instead I use one continuous spline i.e. G3 connection to the flat side of the pill/lozenge (set back from the natural arc tangent point) and normal to the centre plane on the other end of the spline. I then match the spline to appear almost identical to the pure arc....I've seen some people automate this process using an algorithm i.e. to take the guess work out of matching the spline to the arc. I think if you use one continuous spline (rather than a spine joining linked to an arc) you avoid the issue you are having with the sweep method, perhaps? I'm currently working on a button recess channel that resides on a compound A-surface and has a pill/lozenge shape at the end of it. Its been an absolute nightmare to retain G3 lol
All good! Yes, I have built pills/lozenges in the same way as you mention. I think I flip/flop between both ways, as you said, doing it in one results in no surface break along the ends. One good thing I have found with using an arc is that once you have figured out the relief angles on the line and arc, you can repeat at different scales without adjusting anything, which is similar to the algorithm for a single spline. I tried sweeping a line along a deg7+ spline and the sweep still deviated from the path, so I think this is an internal SW issue. Will investigate further. Your current modelling challenge sounds really interesting! I can see how that might be tricky! Have a good weekend.
"SolidWorks must cover up their tolerance issue" When I was starting modeling with Rhinoceros in 1998, there was general opinion that the software was great but the precision tolerances are terrible 😅, which was true,but I don't know how it is now. After a few years of working with Rhino I switched to SolidWorks and just occasionally come back to Rhino to model more complex shapes. I wonder why you are using Solidworks for surface modeling instead of Rhino, is just because surfaces are parametrically controlled? So the style spline between the line and arc is just to assure the G3 constrains? Still a long way for SW to become more efficient in modeling complex shapes, it is likely to be replaced sooner by new technology supported by artificial intelligence.
SW has no user definable modelling tolerance, whereas Rhino does. Well, not that I know of apart from the tolerance setting in knit... Knowing much more than I did in 2001 when trying to send Rhino models via IGES and STEP to engineers using ProE, I suspect I should have tightened up the default tolerances that I had in Rhino! The jump from RH2 to RH3 also helped lots with the exports (like being able to resolidify the parts in ProE). SW must have various tolerances it uses internally, like in this video where the output was not following the input geometry I use Solidworks as I have clients that need models built in it. Also because it is parametric, as you mentioned. It is way easier to change a heavy model in SW than it is in Rhino. Manually manipulating surfaces in Rhino wins hands down, but balanced against having to layer off progress in case I need to change something - SW wins, even if surfacing can be problematic. Yes, the style spline is there to ensure a G3 connection.
beautiful ! thank you!
All good, thanks for watching.
I've got a soft spot for lozenge shapes too! But you probably already knew that! xD
Yeah for sure! Found a picture of a more challenging lozenge form but there’s no info about the origin. Maybe I will model that next time.
@@AndrewJacksonDesignStudio I still feel a perfect g2 curve transition between the line and arc is the best.
Anyway, after lozenges maybe try g3 ball corner cubes! :D
Pretty sure this has smoother curvature transition than a g2 transition. Also have to balance between eroding too much of the arc and having the transition being too abrupt.
G3 and ball is an oxymoron? If the ball implies an arc or circle that is 🤔
@@AndrewJacksonDesignStudio continuing my point above, this perfect g2 “transition” curve I’m talking about also works great when you have to make offsets.
Offsetting any line and arcs with g3 curve between them leads to discontinuities which when you try to correct will result in variations in offset distance between the entities
Really great video as always, cheers A J!
I go about this using a similar method in PTC Creo/Style (ISDX), but I don't use the section of arc for the top profile driver. Instead I use one continuous spline i.e. G3 connection to the flat side of the pill/lozenge (set back from the natural arc tangent point) and normal to the centre plane on the other end of the spline. I then match the spline to appear almost identical to the pure arc....I've seen some people automate this process using an algorithm i.e. to take the guess work out of matching the spline to the arc. I think if you use one continuous spline (rather than a spine joining linked to an arc) you avoid the issue you are having with the sweep method, perhaps?
I'm currently working on a button recess channel that resides on a compound A-surface and has a pill/lozenge shape at the end of it. Its been an absolute nightmare to retain G3 lol
All good! Yes, I have built pills/lozenges in the same way as you mention. I think I flip/flop between both ways, as you said, doing it in one results in no surface break along the ends.
One good thing I have found with using an arc is that once you have figured out the relief angles on the line and arc, you can repeat at different scales without adjusting anything, which is similar to the algorithm for a single spline.
I tried sweeping a line along a deg7+ spline and the sweep still deviated from the path, so I think this is an internal SW issue. Will investigate further.
Your current modelling challenge sounds really interesting! I can see how that might be tricky! Have a good weekend.
Had a crack at some forms on a compound surface. Revolved surface, so not that much to think about :)
th-cam.com/video/Fgli405qRmk/w-d-xo.html
"SolidWorks must cover up their tolerance issue" When I was starting modeling with Rhinoceros in 1998, there was general opinion that the software was great but the precision tolerances are terrible 😅, which was true,but I don't know how it is now. After a few years of working with Rhino I switched to SolidWorks and just occasionally come back to Rhino to model more complex shapes. I wonder why you are using Solidworks for surface modeling instead of Rhino, is just because surfaces are parametrically controlled? So the style spline between the line and arc is just to assure the G3 constrains? Still a long way for SW to become more efficient in modeling complex shapes, it is likely to be replaced sooner by new technology supported by artificial intelligence.
SW has no user definable modelling tolerance, whereas Rhino does. Well, not that I know of apart from the tolerance setting in knit... Knowing much more than I did in 2001 when trying to send Rhino models via IGES and STEP to engineers using ProE, I suspect I should have tightened up the default tolerances that I had in Rhino! The jump from RH2 to RH3 also helped lots with the exports (like being able to resolidify the parts in ProE).
SW must have various tolerances it uses internally, like in this video where the output was not following the input geometry
I use Solidworks as I have clients that need models built in it. Also because it is parametric, as you mentioned. It is way easier to change a heavy model in SW than it is in Rhino. Manually manipulating surfaces in Rhino wins hands down, but balanced against having to layer off progress in case I need to change something - SW wins, even if surfacing can be problematic.
Yes, the style spline is there to ensure a G3 connection.