LDCad—Tying the knot

[N. W. Perry]

Me again. :-)

Any way to get the bezier curves tighter in the string template? The current limit of 5 isn't quite enough to tie this not:    

It looks okay from above, but in the front view, you can see the loop doesn't draw in snugly. (I tried adding an extra path point on the bottom of the loop, but it just made all the corners square off.) I should also add that I'm using the Studio workaround, so the donor parts are 4-4cyli's rather than LDCad's donor parts. (It actually seems to curve more tightly this way.)

[Roland Melkert]

Any way to get the bezier curves tighter in the string template? The current limit of 5 isn't quite enough to tie this not

I'm afraid they are currently hard coded with a minimum of 5.0

I'll lower it to 1.0 for the next version.

[N. W. Perry]

Roland Melkert pid='34568' dateline='1573674204']
I'm afraid they are currently hard coded with a minimum of 5.0

I'll lower it to 1.0 for the next version.

Wow, I actually discovered something beyond the capability of LDCad? 

Does anything dangerous happen if you go below 5, or was it just thought unlikely that anyone would need a lower setting?

I also meant to report, for the part where the string is wrapped around a spool, I did it with bezier points but I assume I should have used circle points…but for some reason I had trouble getting them in line correctly. Still, beziers worked out handsomely:
   

[Philippe Hurbain]

It would also be possible to create a knot part, but I'm not sure it would be better/easier to use...

I also meant to report, for the part where the string is wrapped around a spool, I did it with bezier points but I assume I should have used circle points…but for some reason I had trouble getting them in line correctly. Still, beziers worked out handsomely:

How brave! I generally go with only 1 layer of string... But yes, Bezier is the right CP to use, circle don't work well in this case (where is the wire supposed to leave one circle to go to another? what path in between?)

[N. W. Perry]

It would also be possible to create a knot part, but I'm not sure it would be better/easier to use...

How brave! I generally go with only 1 layer of string... But yes, Bezier is the right CP to use, circle don't work well in this case (where is the wire supposed to leave one circle to go to another? what path in between?)

Well, at least I didn't attempt to tie the knot under the string, where it attaches to the spool! 

So is there an application where it does make sense to mix circle and bezier points? 'Cos that's exactly the problem I had: how to get the string to follow the correct path in and out of the circle.

[Roland Melkert]

So is there an application where it does make sense to mix circle and bezier points?

The circle points only play nice combined with bezier points if the whole thing is on a 2d plane.

If not weird thing will start happening the more a 3rd axis comes into play.

I intended it mostly for guiding bumps in bands (gravity pull etc).

As for spools I'm thinking about a third control point for those so they can be generated based on a single set of parameters (bit like springs).

[N. W. Perry]

The circle points only play nice combined with bezier points if the whole thing is on a 2d plane.

If not weird thing will start happening the more a 3rd axis comes into play.

I intended it mostly for guiding bumps in bands (gravity pull etc).

As for spools I'm thinking about a third control point for those so they can be generated based on a single set of parameters (bit like springs).

I see. Well, I do think it would be handy to have some way to automatically generate spools—and knots, too—since both are used in a lot of official models. I can say that with spools, I was surprised to find that there's some complexity at the end of each course of windings, where the strand crosses over itself and rises up to the next larger diameter. It's not quite a simple as a given number of concentric coils, if you're going for total realism.

(Actually, I find that many of the modeling challenges I face have to do with just figuring out the correct point at which two or more objects would physically collide: place this path point so that the string's outer diameter contacts the inner diameter of this hole…rotate this part so that its end just touches this other part…place this oddly-shaped part on the ground so that its various surfaces come to rest as they naturally wood…and so forth. That's a big part of making a spool, too—not terribly difficult, but fiddly and time consuming. I often wonder if programs could have an overall awareness of the physical mass of the modeled parts, and be able to constrain and automate their placement accordingly—but that's really another thread!)

[Philippe Hurbain]

As for spools I'm thinking about a third control point for those so they can be generated based on a single set of parameters (bit like springs).

Speaking of springs (and spools) it would be nice to have the possibility to create a LQ version, cylinder segments based (smaller size, direct compatibility with Studio renderer)

[Roland Melkert]

Speaking of springs (and spools) it would be nice to have the possibility to create a LQ version, cylinder segments based (smaller size, direct compatibility with Studio renderer)

I think the main reason was generating the mesh was surprisingly easy.

Why Studio only supports type 1 lines in custom parts is beyond me.

[Willy Tschager]

Looks like you have too much spare time at your disposal  

w.

[N. W. Perry]

I think the main reason was generating the mesh was surprisingly easy.

Why Studio only supports type 1 lines in custom parts is beyond me.

Yeah, I would rather have Studio accommodate the standard LDraw editors than the other way around! 

Looks like you have too much spare time at your disposal  

w.

You'd think so…but that right there took me something like a week to finish. 

[Max Martin Richter]

You'd think so…but that right there took me something like a week to finish. 

I do the same on several models. Just look in the OMR for 8856 for example. The trick I use: I just add on round of the rope open a text editor and copy and manipulate the rest of this rope layer by hand. Only hard thing is, when you have to go to the second layer. This is done manually and after the first full rope round in the second layer I use the text editor again.

Doing a spool takes half an our by using the text editor. ;-)

I have the bad feeling that I have to "borrow" your knot for the next model. :-D

/Max

[N. W. Perry]

I do the same on several models. Just look in the OMR for 8856 for example. The trick I use: I just add on round of the rope open a text editor and copy and manipulate the rest of this rope layer by hand. Only hard thing is, when you have to go to the second layer. This is done manually and after the first full rope round in the second layer I use the text editor again.

Doing a spool takes half an our by using the text editor. ;-)

I have the bad feeling that I have to "borrow" your knot for the next model. :-D

/Max

Well, I'll make a better one perhaps, whenever the next version of LDCad appears. Then I'll be happy to offer it up as a template. 

[Roland Melkert]

Well, I'll make a better one perhaps, whenever the next version of LDCad appears. Then I'll be happy to offer it up as a template. 

If you take the generated code (stuff below 0 !LDCAD GENERATED) and put it in a custom part file, you could use it as a donor inside a second rope template (which would use it as a cap).

[N. W. Perry]

If you take the generated code (stuff below 0 !LDCAD GENERATED) and put it in a custom part file, you could use it as a donor inside a second rope template (which would use it as a cap).

Neat idea—and it won't matter that the knot already contains a cap part? (I added the 4-4 disc that wasn't present in the LQ rope template, at least for the visible end of the string.)