Smooth (tangent) transition from a straight path to a helix path

milan_jovanovic

I am trying to get a smooth transition from a straight section to a helix. Like in a torsional spring, for example.

To do that, I am preparing a multi-part path, so that I can then sweep along it. I encounter two problems:

• I don't know how to fix the helix starting point to the end point of the straight path. I can get very close by changing the start angle of the helix.
• But then the helix path I get starts vertically at an angle to the straight path (it does not start as a tangent).

How can this be done properly?

mahir

It feels a bit messy, but you can use a bridging curve to generate a point to point tangent line. You'll have to create a couple planes - one normal to the helix at an endpoint and another parallel plane offset by the desired length of the straight section. Then, create a sketch on the offset plane that has a single point converted from the helix endpoint. By construction, the line connecting the helix endpoint and this sketch point will be tangent to the helix itself. In my example I controlled the angle between start and end points by inputting an appropriate number of turns. For example, anything ending in .5 will be parallel (0 deg apart). Anything ending in .25 will be 90deg apart. In my example I chose 3.375 deg in order to get a 45deg angle.

https://cad.onshape.com/documents/57acdfaae4b005c413ed9b6f/w/3fd585a46d3af1b3ba413c53/e/1fff1f27f6553a553baa7ad0

alnis

Another way is to create planes on the ends of the coil, create sketches using on-the-fly mate connectors, realign them with those planes, rotate them on the y axis 90 degrees, and use arcs for a smooth tangent transition:
https://cad.onshape.com/documents/d52274d79094d1ad5ba30a71/w/184b8bd272091790e4a2c6c7/e/066aba2f72156eb326a1f29d

This seems pretty stable, but it's not very intuitive and maybe has some big problems I'm not seeing right now.

mahir

It feels a bit messy, but you can use a bridging curve to generate a point to point tangent line. You'll have to create a couple planes - one normal to the helix at an endpoint and another parallel plane offset by the desired length of the straight section. Then, create a sketch on the offset plane that has a single point converted from the helix endpoint. By construction, the line connecting the helix endpoint and this sketch point will be tangent to the helix itself. In my example I controlled the angle between start and end points by inputting an appropriate number of turns. For example, anything ending in .5 will be parallel (0 deg apart). Anything ending in .25 will be 90deg apart. In my example I chose 3.375 deg in order to get a 45deg angle.

https://cad.onshape.com/documents/57acdfaae4b005c413ed9b6f/w/3fd585a46d3af1b3ba413c53/e/1fff1f27f6553a553baa7ad0

alnis

Another way is to create planes on the ends of the coil, create sketches using on-the-fly mate connectors, realign them with those planes, rotate them on the y axis 90 degrees, and use arcs for a smooth tangent transition:
https://cad.onshape.com/documents/d52274d79094d1ad5ba30a71/w/184b8bd272091790e4a2c6c7/e/066aba2f72156eb326a1f29d

This seems pretty stable, but it's not very intuitive and maybe has some big problems I'm not seeing right now.

milan_jovanovic

Thank you very much @mahir and @alnis! I think I will be able to use any of these solutions.

One issue might arise. This way straight paths are defined from the helix path. So the whole, more complex part, must be defined starting from the helix. For some parts that may be impossible.

mahir

milan_jovanovic said:

One issue might arise. This way straight paths are defined from the helix path. So the whole, more complex part, must be defined starting from the helix. For some parts that may be impossible.

If it's important that the helix be driven by the straight sections instead of the other way around, then there are ways to do that. But again, it's a bit messy/complicated. Just like I manually backed out the correct number of turns for a particular angle, this could be automated via equations, variables, and/or featurescript. In this second example I laid out two arbitrary line segments. The location of the helix is driven by these segments. While the construction is probably not robust enough to withstand the angles changing 180deg, it is relatively flexible. Given the desired helix hight, number of full turns, and the axial blend distance, the helix is then generated by measuring the line segment angles.

https://cad.onshape.com/documents/57acdfaae4b005c413ed9b6f/w/3fd585a46d3af1b3ba413c53/e/c2625bd637f213bfc63464c3

milan_jovanovic

I will need some time to try this out.

Thanks @mahir!