Smooth termination of bottle thread?

laird_broadfield

(Disclaimer: amateur.)

I've correctly profiled the SPI-415 thread on the mouth of my bottle (helix, curve-point plane at the end of the helix, sketch the profile, sweep.)

That's here: https://cad.onshape.com/documents/8804c0b6e8623410630d22f0/w/654293064dc85d3482636805/e/1cada1b1e0742069abf96827



However, it ends abruptly, when it should taper down smoothly.  The SP-415 standard doesn't actually specify the form of the taper, but specified or not, I'm not really sure of the best way to model it.

SP-415 thread drawing (https://www.bevtech.org/assets/Threadspecs/sp415.pdf)


I've considered
 - lofting the end profile down to a point, but that doesn't want to cooperate without a bunch of guides and still isn't elegant.   
 - adding a path that dives into the surface, but that doesn't achieve a taper.  
 - applying some form of radius to the end, but that doesn't cooperate very well either.

Thanks.

robert_morris

@laird_broadfield,

When I had to do this in Solidworks, I created another helical path at the end of the thread that was tangent to the thread path and dove into the surface. I then swept the profile along this path. Once the profile is swept, it will naturally create a taper. See the screenshots below.

brucebartlett

In the past, I recall just revolved the end face back into the main body. However, you might have to put a sketch on the end face to get an axis. 

laird_broadfield

robert_morris said:
When I had to do this in Solidworks, I created another helical path at the end of the thread that was tangent to the thread path and dove into the surface. I then swept the profile along this path. Once the profile is swept, it will naturally create a taper. See the screenshots below.

Thanks -- I tried that at some point, and it works, and as a practical matter it probably works just fine -- but it didn't seem very elegant:



If I clean up the extra extrusion and ignore the (microscopic) edges and corners, it's not bad, though:



Just wondered if there was a better approach with conical fillets or something.

robert_morris

@laird_broadfield ,

I haven't looked at you document yet, but one thing you could try and do is move the fillet at the base of your thread to be applied after you do the sweep on the ends. That will help blend them in to the bottle neck. Another thing that I see in your top image is you should extend your curve so that it goes inside of the bottle surface, not just up to it. That will get rid of the extra little flat spot on the end of the thread.

Hope that helps,

laird_broadfield

robert_morris said:

@laird_broadfield ,

I haven't looked at you document yet, but one thing you could try and do is move the fillet at the base of your thread to be applied after you do the sweep on the ends. That will help blend them in to the bottle neck. Another thing that I see in your top image is you should extend your curve so that it goes inside of the bottle surface, not just up to it. That will get rid of the extra little flat spot on the end of the thread.

Hope that helps,

Hmm interesting.  I don't have a fillet feature at the base; I sketched it in from the spec.  I didn't think of sketching it unfilleted, and filleting afterwards, though; that's great.  Thanks.

laird_broadfield

MUCH better.  Thanks.

NeilCooke

@laird_broadfield long time since I made one of these, but in your revolve sketch, put the revolve axis further away from the thread profile. That will give you a larger more elegant blend. 

laird_broadfield

NeilCooke said:

@laird_broadfield long time since I made one of these, but in your revolve sketch, put the revolve axis further away from the thread profile. That will give you a larger more elegant blend. 

Okay, I want to understand that, and I don't.  I have no revolves in my document; I have a sweep along the helix, and I have a sweep at each end that extends the profile along a bezier beginning at the top center of the profile.  This is from the tab "after forum feedback" in the above document:



When you say "larger more elegant blend", do you mean the thread completion segment's dive into the body?  If so, wouldn't I just extend the bezier and tweak the handles?  I could do that with a revolve instead of a sweep, I suppose.  It didn't occur to me to use a revolve, partly because I can't see an easy path to replicating the curve up to the other end via dependencies -- is there a reason a revolve would be better?

If you mean the thread profile's radius into the body, that's 0.5 in the SP-415 spec.

NeilCooke

Sorry, I commented without looking at the doc. Here's how I always did it (not perfect but near enough). See last 3 features.

https://cad.onshape.com/documents/14ac901b774df8920196fd26/w/78cd94c4e521567425b1f986/e/657f33601d72149d48f6890d

laird_broadfield

Well, I tried a whole 'nother approach.  I'm not sure if it's better, or just different.  So far, all of the solutions dive into the surface, but they really only round off due to the taper of thread; they're not really getting smaller as they go.  Here's a different technique (on top; old method below; in document tab "after after forum feedback"):



I used Offset Surface to create a copy of the surface of the flat end of the thread.  Then, two transforms to move that surface to the end of the spline and to scale it to 30%, then a loft from the end of the thread to the scaled down surface.  Thoughts?

ottmar_brandau

Hi,

Could you be more specific on how to do the offset surface and transform commands. When I offset it the resulting geometry is not where I need it to be to do a loft later.
Thanks.

ottmar_brandau