Christiaan Posted February 15, 2007 Author Share Posted February 15, 2007 Couldn't be more serious Gytis. My reputation as kingpin CADster in my office depends heavily on the gurus of this forum. Quote Link to comment
Abacus Posted February 15, 2007 Share Posted February 15, 2007 Christiaan, Try it using Arcs instead of circles. The fillet tool does the trimming. ? Arc by radius, make approx. 300 degree sweep ? Duplicate/Mirror (openings in arcs face each other) ? Adjust radius in second arc. ? Fillet w/ trim (touch arc ends) Quote Link to comment
propstuff Posted February 15, 2007 Share Posted February 15, 2007 Do you gentlemen have jobs...or is it as for me, this forum is just more fun? Geometry "FUN"?? You would make Euclid turn in his grave! Gytis: The Harmony of the Spheres is a very, very spiritual matter, and not to be trifled with! ;-) Besides which, if you're designing "product" rather than buildings, tangents to circles is a fundamental part of life. It's also something where the VW -sometimes-I'll-work-and-sometimes-I-won't tangent snaps are a real pain. cheers, Quote Link to comment
Benson Shaw Posted February 15, 2007 Share Posted February 15, 2007 To estimate your fillet radius, draw a 3 point arc between the circles which sort of looks tangent to both circles and read its radius in the OIP. Then in the fillet tool assign a fillet radius value close to your estimate and create the fillet. Adjust the fillet radius in the fillet tool and create new fillets until you get the look you are after. Or use Mike m Oz's method: draw a circle or arc that looks about right, drag it close and use constraints or parametric constraints to place it tangent to both. Trim. It would be great if you could use parametric tangent constraints to keep the fillet tangent to both circles, and then move the arc center or change its radius in the OIP to adjust the fillet config. Unfortunately, VW moves the circles and the arc to keep everything tangent. Adding parametric distance constaint between the circles to keep them in place crashes VW when arc radius is adjusted (well, the beach ball of doom spun longer than my patience allowed). Locking the circles to keep them in place does not work either. VW dissallows changing radius of tangent constrained arc, and will not allow constraints on locked objects. So where is the center of any tangent arc/fillet? Draw a bunch of fillets and note that chords of all these fillets are not parallel. They sort of fan out. Therefore, the centers of all possible fillets lie on a curve. Some smarty pants out there can give us a mathematical description of that curve (not me!), if it seems important to prelocate the center. Me, I'm just glad VW can calculate the fillets tangent to both circles. And, yes please, could we get that tangent constraint and tangent arc tool fixed so we can make arcs tangent to objects at both ends during creation. -B Quote Link to comment
mike m oz Posted February 15, 2007 Share Posted February 15, 2007 Christiaan, try reading the instructions again - it really is quite simple. Your two circles are circles A and B. The fillet arc is circle C - as you apply the constraints circle C will move to match the constraint conditions. Once you have the three circles constrained together you can adjust the size of circle C by dragging on its perimeter handle. Be warned though that it is now circle C which will stay still and circles A an B will both move. I trust you do inform others in your office where you have obtained many of the answers to their questions from, and that you are not just using our goodwill for your persona gain. Quote Link to comment
P Retondo Posted February 16, 2007 Share Posted February 16, 2007 Christiaan, why sweat it? Let VW do the geometry. For another piddly 11 seconds, you can use the fillet tool, which automatically creates fillet arcs that are tangent to both circles. You can set the radius of the fillets to whatever value is within range. If you have other constraints, then maybe you'd have to go the hard way and construct a custom arc. As pointed out, the 3 point arc tool can give you something approaching tangency at the intersection points, but will not be exact unless you are incredibly lucky. Quote Link to comment
mike m oz Posted February 16, 2007 Share Posted February 16, 2007 Christiaan - I think Pete wins hands down. The big question though is what you are going to reward him with? 50 Euros sound about right. Quote Link to comment
Christiaan Posted February 16, 2007 Author Share Posted February 16, 2007 I reckon your circle method's pretty good Mike. Quote Link to comment
Christiaan Posted February 16, 2007 Author Share Posted February 16, 2007 And the fillet method from Mike. Quote Link to comment
islandmon Posted February 16, 2007 Share Posted February 16, 2007 Here's a trig solution based on the Ratio of 10 : Phi showing the specific Arc & Line relationships mention previously : Quote Link to comment
P Retondo Posted February 16, 2007 Share Posted February 16, 2007 Christiaan - I think Pete wins hands down. The big question though is what you are going to reward him with? 50 Euros sound about right. Christiaan's help to everyone on this forum gives him plenty of credit with me! Islandmon, I believe that for two arcs to be mutually tangent, the two centers and the point of coincidence have to lie on a line. Your lower tangent point doesn't seem to meet that requirement. Quote Link to comment
propstuff Posted February 16, 2007 Share Posted February 16, 2007 Islandmon, I believe that for two arcs to be mutually tangent, the two centers and the point of coincidence have to lie on a line. Your lower tangent point doesn't seem to meet that requirement. Yes indeed they do: by definition they must share radii/normals to be tangential. N. Quote Link to comment
islandmon Posted February 16, 2007 Share Posted February 16, 2007 So right you are , Pete, the arc was correct @ 104.4746? ... but the arc dimension @ 104.51719? on the image was incorrect ! I posted a revised image. Quote Link to comment
islandmon Posted February 17, 2007 Share Posted February 17, 2007 Just for fun ... here's a pure tangent shape based on sqr5: Quote Link to comment
Christiaan Posted February 17, 2007 Author Share Posted February 17, 2007 It looks edible. Quote Link to comment
islandmon Posted February 19, 2007 Share Posted February 19, 2007 This is my 'Best Fit' solution to Christiann's challenge: Quote Link to comment
quigley Posted February 19, 2007 Share Posted February 19, 2007 My vote is for the 2 circles, 2 fillets, trim and compose approach :-) But......if you want to be able to edit the fillets (ie the tangent arcs) or indeed the circles the only way to do this is to use geometrical constraints...which VW can do....sort of. Quote Link to comment
Christiaan Posted February 19, 2007 Author Share Posted February 19, 2007 I've found the method of estimating the radius with a circle then 2 Fillets, Trim, Compose to be the quickest method. 26 seconds to be precise. So thanks for everyone's input; it'll help with many other shapes too I'm sure. By the way, does anyone know the answer to my query about the advantages and disadvantages of converting it into a polygon? One disadvantage is that you lose the snap centres of the circles. Quote Link to comment
mike m oz Posted February 19, 2007 Share Posted February 19, 2007 Christiaan - you have answered your own question. Polygons are just shapes made up of short lines. Far harder to deal with. Quote Link to comment
MullinRJ Posted February 20, 2007 Share Posted February 20, 2007 I can't for the life of me figure out what I'm doing wrong though. When I get to the point where I draw the arc it never meets the assumed centre line (see dotted line)?? Christiaan, It is because the center of the arc is only on your assumed center line when the radius of the arc is infinite. Notice you started with a line tangential to both circles. That is equivalent to an arc of infinite radius. When the radius is less than infinite, the center point of the arc falls to one side of your assumed center line, namely, it falls to the side of the smaller circle. The path of all possible centers traces an hyperbola and one asymptote is the line you drew. The exception is when both circles are of equal radii, then the expected path of center points is vertical and bisects the line connecting the two closest points on the circles. Plotting the family of possible curves is quite beautiful. Perhaps Islandmon will post a picture if I send him the scratch file I've been using. Raymond Quote Link to comment
Petri Posted February 20, 2007 Share Posted February 20, 2007 Geometry is that which, even if you ignore it, does not go away. Whilst all progress in the world depends on the Unreasonable Man, even he can't change the way geometry works. Quote Link to comment
Christiaan Posted February 20, 2007 Author Share Posted February 20, 2007 Christiaan - you have answered your own question. Polygons are just shapes made up of short lines. Far harder to deal with. So is there any advantage to a polygon over a closed polyline? Quote Link to comment
islandmon Posted February 20, 2007 Share Posted February 20, 2007 Plotting the family of possible curves is quite beautiful. Perhaps Islandmon will post a picture if I send him the scratch file I've been using. sure nuf ... send: ejarmstrong@gmail.com Quote Link to comment
islandmon Posted February 22, 2007 Share Posted February 22, 2007 RJ Mullin sent me a file with 20 Tangent Curves to play with : http://www.fractalnet.org/gallery2/v/TangentCurves/ Quote Link to comment
MullinRJ Posted February 22, 2007 Share Posted February 22, 2007 ej, thank you. You've made it look exceptional. The renderings in your gallery - even more so. Christiaan, One thing that did not show up in the picture above are the markers along the green line that correspond to the centers of the displayed fillets. The centers of all possible fillets lie on the green line and the green line gets closer to the red asymptote as the fillet radius increases, but never crosses it. In the file I sent ej the fillet centers were marked with Loci, which obviously did not print/render. I was expecting a screen shot, but I got a work of art. I really should have expected this, considering the source. Thanks again, ej. Raymond Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.