Unsegmented Faces on Swept Custom Thread

[afgoody]

Following the example provided by @HughesTooling here, I'm trying to get unsegmented faces on a swept custom thread, but cannot seem to get it to work. Any suggestions?

[HughesTooling]

@afgoody You just need to swap your selections for path and guide

[afgoody]

Well, dang it! Rookie mistake! Thanks and sorry for wasting your time.

[andrew2085]

@HughesTooling I'm not sure it matters in this specific situation, but wouldn't this give the incorrect thread geometry? Thread profiles are usually specified in the axial plane, not the plane perpendicular to the helix. If sweep works better with a plane perpendicular to the helix (does it actually?) then wouldn't you need to sketch the thread profile in an axial plane then project it to the helix perpendicular plane?

[afgoody]

@andrew2085 I thought about this before trying it out. His method specifies the cross-sectional profile geometry only in the perpendicular plane. The rest of the thread's geometry (i.e. pitch, length, revolutions) is still specified in the coil. And yes, it does work so much better/cleaner/

[andrew2085]

I looked a bit more into this and the thread profile you get with a sketch perpendicular to the path is wrong for even more reasons. A sweep with a sketch on the axial plane comes out perfect. This happens because the part of the profile closer to the center lags behind the part on the outside (on the top, leads on the bottom), by the time the inner part of the profile reaches where the outer side just swept, the profile has moved further up the helix, leaving a curved thread flank.

With a triangle swept perpendicular to the helix:

With the same triangle swept from the axial plane:

I used an extreme helix to make the effect visible to the eye, but it will always be there even with more normal threads. The projected thread profile in the first method will always produce splines on the flanks, and the second will produce lines.

[afgoody]

@andrew2085 Interesting. Care to share your test files?

It seems clear to me that Autodesk needs to expand their coil feature to support custom profiles. Until then, I suppose I’m still on the search for the best workaround.

[andrew2085]

It depends what you're modeling. If you're modeling e.g. a bar bent into a helix, then the method you did with the sketch plane perpendicular to the helix is correct. If you're modeling a thread that has the thread profile specified on the axial plane, the you should use a sketch on the axial plane. The segmenting you were getting earlier were from mixing up the path and guide curve, not the profile sketch orientation. Both plane orientations produce clean, but different geometry.

[andrew2085]

Sure here's an example. Also it's worth noting that the coil feature sweeps a profile on the axial plane, not the helix perpendicular plane, so it probably should have been named differently as you don't get the geometry you would get if you e.g. took a round bar and bent it into a coil, it has the geometry you would want for a thread.

[johnsonshiue]

Hi! To obtain even higher precision, you may want to use Solid Sweep to create the thread.

Many thanks!

[andrew2085]

Thanks for the tip. Out of curiosity, on a technical level why does this lead to a more accurate result?

Edit: Ahh I see there is an Aligned option to choose the helix axis. Not sure if that is the reason why though, or if there is more too it.

[johnsonshiue]

Hi Andrew,

Solid Sweep offers higher geometric precision, because it mimics the actual machining process using a 3D tool to cut. It also means the resultant groove allows the 3D tool to pass through smoothly. While, the Profile Sweep only guarantees the profile to pass through.

Many thanks!

[afgoody]

@andrew2085 Thanks! I tried your example and the differences between an axial and helix aligned profile are much more pronounced with such a tight coil - less so above ~1 mm. But, it is definitely better at least from a best practice perspective in that the profile remains axially-aligned and can still result in continuous faces. So, I implemented that approach in the design I'm working on. It seems the commonality to avoid non-continuous faces is to use the center axis as the path and the coil as the guide rail also avoiding the use of a guide surface altogether. The trick to this method of course is ensuring the path and guide rails are aligned and dimensioned properly so the sweep can use them - a little tricker when the coil starts somewhere other than an origin plane. I think I got the hang of it now though.

@johnsonshiue I gave it a shot. Since it needs a body to sweep with, I just extruded the same shape as the other two methods and tried to use that, but alas...

What did I do wrong? Is this Solid Sweep meant to be Autodesk's answer to requests for a custom threads capability or is there any chance of just adding custom profiles to the coil and/or thread command? It seems custom threads should be configurable straight in coil/thread without requiring all of these additional preparatory steps.

[andrew2085]

I think you just need to add a draft angle on the extrude for your triangle. Basically so only the leading face would leave a trail if you get what I mean. I would expect it to work either way as it kind of defeats the purpose of a solid sweep if it doesn't. I think its just kind of fragile.

Edit: Also about your mention of not needing a guide surface, you can actually use the top plane as the guide surface in this example and it has the same effect as the helix axis as the path.

[afgoody]

Yeah, I think all of these methods to get a custom thread are unnecessarily complex and fragile. I tried the draft and then solid sweep and it still wouldn't work. File attached.

[andrew2085]

I also wasn't able to get it to work here for some reason. The solid sweep seems to be very fragile. I think it has to do the the type of degree-3 b-spline representing that edge having small discontinuities in the curvature. If you generate a helical degree-5 control point spline it works much better, but is not parametric.

[afgoody]

Yeah, see @johnsonshiue, we need a better custom thread tool. Thanks!

[andrew2085]

---

@afgoody wrote:

Yeah, see @johnsonshiue, we need a better custom thread tool. Thanks!

---

We don't need a custom thread tool, and we didn't need a coil tool. We need a helix curve like basically every other CAD program has. Strong fundamentals are much more important than bells and whistles. Once you have a helix primitive you already get a custom coil and thread tool by extension because the helix and sketch could be driven parametrically to create any custom coil or thread you want.

[afgoody]

I don’t know about any other CAD or their helix tools. All I’m saying is we need a less cludgy way to make custom threads. They can easily package and call it whatever they want as long as it works well and in fewer operations.

[andrew2085]

---

@afgoody wrote:

I don’t know about any other CAD or their helix tools. All I’m saying is we need a less cludgy way to make custom threads. They can easily package and call it whatever they want as long as it works well and in fewer operations.

---

You're right. The existing hole and thread features just read major, minor, and pitch diameters from XML files so there is no reason the UI shouldn't also allow entering custom values into text inputs manually. That's very low hanging fruit. It would be even better if, in addition to that, they allowed selecting a sketch profile to use as the thread profile. There's more you can do with a helix than just create threads though, so I still think that should be available.