Hi Larry:

In this case, the minimum bending radius would be 2/3 the thickness of the wire = 1/3 = 0.33 x 2 = 0.66 mm, if you go below that radius, Sweep does not work.
I did it with a radius of 0.75 applying the safe formula that is (thickness * 1.5) / 2, 1 x 1.5 = 1.5 / 2 = 0.75 mm, it is the most used in processes of bending tubes or round bars.

cheers

Thank you for your feedback Booleano.
Our exchanges allow progress. If I offended you by my remark about the fact that the spring must not bite into the wood, I want to tell you that I had not downloaded your file when I wrote this. but if that was the case, I apologize to you.

Well, that said, for me too sweep (tuber in French) did not work (I had aberrations). By using "ReconstructCurve" and playing on the number of points, Sweep works without changing the scale (well for me of course!). Your suggestion for a change of scale (x10) is interesting, maybe a MoI3D expert could give us some information on this subject.
I find that the two scripts "Rebuild" and "ReconstructCurve" are very interesting and very good at solving problems of aberrations during casing (Sweep), among others.

> this would be wrong in the real world. A teel wire would break when bending.
I accept with humility your remark which is well founded.

Boileau:« Hâtez-vous lentement et, sans perdre courage,
Vingt fois sur le métier remettez votre ouvrage :
Polissez-le sans cesse et le repolissez… »

"Hurry slowly and, without losing heart,
Twenty times on the loom, hand over your work:
Polish it constantly and repolish it… ”

- Zooen

Hi Zooen.

You do not have to apologize for anything, I have not felt offended at all, because it is true that the first model that I made the dock if it bit the wood.
This forum is simply amazing, I follow everything that is published here, because it is how I learn every day to use the best methods, and I have never read a post or response with bad words, criticisms, offenses or reports, I only read respect, collaboration , help and solutions to almost everything and in record time.
The scale technique has been advised on several occasions by Michael, who is the highest authority on this forum.

Cheers.

Re,
I apply the quote from "Boileau".

By observing very closely the curve of my spring, I noticed that it had misalignments of control points (it is perhaps a lack of control of "ReconstructCurve").
So I went back to my job, it's a little better.
I know that the curve of the small angle is still not compliant for a fabrication (machining). But hey, I trust the technician who will be able to remedy this detail!

I would appreciate that an expert (there are some who come to the forum) explain to us the right method to achieve this kind of spring.

And Booleano inspired me, I made a better quality Blender rendering (again I should work a little more on the subject ...)

- Zooen

Hi Zooen

Here you have 2 videos in both cases they use a 1 mm bending radius, looser than the 0.75 mm that I use, and in both cases the Helix outputs are tangent.
And this is the only problem you have.

Cheers


https://www.youtube.com/watch?v=ug7bMScWiYk&ab_channel=SolidWorksTutorial%E2%98%BA
https://www.youtube.com/watch?v=J07X3WAdn7g&ab_channel=EtienneK

Remember that you can also use the Blend function on opened cylinders! ;)

Thanks Booleano for these links to the videos.

I don't know it's videos. I had no idea that the clothespin inspired other people on other software.
After all, it's okay to do it yourself, to make mistakes.
It is often through our mistakes that we progress, if we have the will to exploit them.

-Zooen

Hi Pilou.

In the virtual world anything goes, and that solution is perfect.
In the real world, to bend these types of shapes, made with Blend, it would be necessary to use hot molding techniques, subjecting the metal wire to a very high temperature to be able to deform the material without breaking it, as is used in forging work.
In the virtual world you can bend a tube or bar at 90º, but in the real world it is impossible to do it without cutting the part into 2 parts with an angle of 45º and then welding them together.
In addition to making springs or springs, steel material is used, not iron, because after giving it the appropriate shape it must be subjected to a tempering process to give it hardness and that it cannot be bent, and with a bending with a constriction lower than minimum radius established by mathematical formulas, the material would break safely.

But I thought I understood that Zooen wanted a method that could be manufactured in the real world, at least that is what I understood about this phrase of his:
"I know that the curve of the small angle is still not compliant for a fabrication (machining). But hey, I trust the technician who will be able to remedy this detail!"

But this type of piece has helped me to verify many undesirable effects of the Sweep tool that I do not know why they occur, and that I need to consult so that someone can explain it to me.

Cheers

Hi Zooen.

I looked it up to see if someone had done it in other software and see what technique they used, and see what bending radius they used out of curiosity.

I have made a video to show the errors, which I do not know why they occur taking advantage of the clothespin spring in this post,

And I would like Michael, or someone who knows why they occur, to explain to me why this happens, why I made the curves following the correct geometry rules, or at least I think so.

The results are absolutely baffling.

Attached 3dm file.

Cheers

Hi Booleano, when you have a joined curve, sweep will try to do mitered corners.

That's done by extending surfaces and then intersecting them with each other which becomes a difficult calculation if your curves are close to being tangent to each other but are off by a little bit as is the case in your curves. When you have curves like that the extensions do not have a crisp intersection.

In your case here the problem is that these curves segments are not tangent to each other, they have a shallow approx 3 degree deviation in their tangents:



- Michael

Like you Booleano, using "Sweep" I had results that were absolutely baffling. I did not know the change of scale (X10) which solves the problem (well, if I understood correctly). So I used "ReconstructCurve" and I didn't like the disconcerting result (which I called aberration). Then, following our discussions, I re-examined my spring by zooming in, I noticed that the control points were no longer aligned in several places. So I did "pincer_02", not perfect, but better.

@Michael> In your case here the problem is that these curve segments are not tangent to each other, they have a shallow deviation of about 3 degrees in their tangents.
But how do you go about drawing, correctly, a curve tangent to the end of the curve of a helix or of a curve in general?

-Zooen

<< tangent ?...
Blend seems makes connection tangent no?

Hi Zooen,

re:
> But how do you go about drawing, correctly, a curve tangent to the end of the curve of a helix or of a curve in general?

Well there are many different ways depending on the particular situation and if you're drawing a line, an arc, or a freeform curve.

Probably in this case if you wanted the rest of the pieces to be all squared up it would probably be good to cut away some space between the helix and the line and put in a blend curve.

The basic problem is that you've got these areas that you want squared up:



But the helix part is a non-planar curve with a curve tangent that goes in a sloped direction like this:



So if you have just a 2D sketch and the helix touching each other there will be a tangent break at the spot where they touch. Putting in a blend curve as a transition between them could work or another possibility would be to edit the first interior control point of the helix to make it in line with the 2D line.

- Michael

Thanks Michael,

I will study these explanations and suggestions.

- Zooen

Hi Michael.
"Putting in a blend curve as a transition between them could work or another possibility would be to edit the first interior control point of the helix to make it in line with the 2D line."

I have not been able to make a fillet curve that functions.
Could you post how it can be done

Thanks

I have made a video with a solution that works respecting the radios.
If there is another easier solution, I would like to know the process.

Cheers

Hi Booleano,

re:
> I have not been able to make a fillet curve that functions.
> Could you post how it can be done

Here's a demonstration:



- Michael

Hi Booleano,

re:
> I have made a video with a solution that works respecting the radios.
> If there is another easier solution, I would like to know the process.

After you cut the curves up, discard the small curve fragment and use Construct > Blend to put in a blend curve, then join those pieces together.

- Michael

Hi Micharl.

It's perfect, I didn't remember that Blend could also be used in curves.
Is there a script to be able to move a line tangent to the perimeter of a circle, ellipse or curve?
In the case of this spring, it is very difficult to place the end 'A' tangent to Helix at point 'B' or at another point.

Move the lines from points 1-2-3-4 around the circle keeping the tangent lines, as seen in the attached image.
I hope the translation is understood

Cheers

Hi Booleano, I'm not sure that i understand your script request. Why do you need to move the lines around, why not just draw a new tangent line at the spot that you want it?

You can move one tangent line around keeping it tangent using the Transform > Orient command, when snapped onto a curve the orientation picker aligns to the curve tangent.

- Michael