Threads? Which one do I use, and are there tricks to getting this right? (and what about tolerances?)

[OceanHydroAU]

I'm modelling a custom nozzle to screw onto my silicone caulking tube [long story - I need hundreds of specific-shaped blobs squeezed out over (to waterproof) old roof-bolts]... and ran into this head-scratcher:-

My nozzle needs an internal female thread to affix to the tube:-

I can measure all those bits with my digital callipers, and in the past when I made a Voss-to-Sodastram adapter, I manually recreated everything from first-principals... which worked great... but took a days to get right!

What is the real way that pro's do this?  How do I know which of all those things to use, and how to determine the parameters "Thread Type", "Size", "Designation", "Class" and if I need to click the "Modelled" box?

I'm guessing...

a) that almost all those Thread-Type options are obscure things I can ignore (And, LOL, "ACME"? Did Wile E. Coyote give them some kind of trademark concession?)

b) I can probably measure the distance of some things to work out one or more of those other parameters

Also - what's the story about tolerances?  I always engineer in a small "gap" (parametric value) when making stuff to join with other stuff, which I find makes everything fit easily, and can be tweaked if I want it more or less snug.  How do we pick (internal female) threads such that it's actually going to fit properly (onto the male tube nipple thread - and for interests sake - if I wanted to 3D print the male thread too - what's the "pro" way to specify that gap)?

In the case of a nozzle thread like this though - I expect I really need something more like a pipe thread?  One which has a slight taper to it, so it connects easily, but fits tighter the harder you screw it down? 

[etfrench]

3d printed threads usually need added clearance that is specific to the printer.  There are too many variables in FDM printers to make generic threads which are accurate.  I usually model the thread, then offset the flanks and top of the thread form.  Print the thread, test fit, and repeat until I get a satisfactory fit.

If you have a flat bed scanner, it should be fairly easy to import a scan to get the thread form.

[OceanHydroAU]

Yep - like I said - I always add a parametric gap variable, so I can tweak that.  For my Anycubic Chiron FFF, the gap is usually the same for everything (big and small, holes/squares/seams/threads*) .. where "*" means I manually made them myself using coils for the path...

I'm hoping that someone who does CAD for a living might read this post and tell me what their thread workflow is... the huge range of unexplained options in there is too bewildering for me (and probably most others I expect).  The *concept* isn't too hard, but it looks like the "standards committee" has been allowed to hold too many meetings...

 

[OceanHydroAU]

What I've learned so far:

1. the world is awash with a dozen different "standards" in just the pipe-industry alone:

https://www.thehopegroup.com/blog/2020/05/learn-all-about-different-types-of-threads/

2. Sometimes we need to use the "hole" command instead of a thread:

https://forums.autodesk.com/t5/fusion-360-design-validate/npt-pipe-threads-new-hole-wizard/m-p/79513...

3.  It looks like the rest of the threads-world (besides pipes) is probably similar to pipes (a dozen different standards) - I've not looked at those yet (since my nozzle seems more "pipe like" than not).

4. Nowhere yet have I found any explanation about tolerances - I presume male will be "smaller" than "female" by some tiny amount (or female will be bigger, or both) - but which is which isn't mentioned - "measuring" to get the size doesn't say "add X for internal" or "subtract X for external" (or both), so I presume when a box-of-bolts says "Size Y" on them someplace, that it's a "white lie", and one or both of the male bolt and/or female nut will never actually be *exactly* that size...

[TheCADWhisperer]

Search Wikipedia on thread nomenclature.

Carefully examine the diagrams.

Now, take a better photo (looking directly at the thread like in the diagrams, not at an angle).

Attach your new picture here.  (I thought I might have a similar tube of sealant laying around but I couldn’t find it for the purposes of this discussion.)

[MRWakefield]

You're correct: there are a lot of thread standards! Some have been superseded or are very seldom used and some are for very specialist applications. ACME (Wil E. Coyote's favourite) for instance is usually used for transmitting power and you wouldn't find it on a bolt.  There are also glass bottle threads and plastic bottle threads and I expect a whole load of others I'm not aware of. You are also correct in that in most cases the male threaded part is smaller than the female. There are standards laid down for all the various thread designations so we as designers/engineers don't generally have to work this stuff out, we just look it up in the relevant standard. You'll find more info if you do a search for thread classes, tolerances and allowances. You should find plenty to keep you occupied if you really want to learn this stuff! 😉

I don't know which thread these sealant cartridges have and without a more orthogonal view it's hard to see the thread form. If I get a moment I'll try to measure one of mine. I probably wouldn't worry too much about the actual standard, I'd just model it up as close as possible to what you've got and add an appropriate amount of clearance, and if you're 3D printing then add a bit more to allow for inaccuracies in the process.

[OceanHydroAU]

Thanks @MRWakefield - very informative.  Good observation on the "glass bottle" threads - I noticed that when making my Voss-Sodastream adapter (  https://a360.co/3687Fbf  ) there is lots of clever stuff they do in there (gas-pressure release vents, funky profiles on the threads themselves, clever ends on both that make mating work nicely, dual-concentric thread-paths, sealing-ends, etc) which both those manufacturers do, and each completely differently...

I love everything "parametric"... so I thought: OK - I'll design a "general purpose" adapter with parameters, so I can share it, which means others can then use this sucker and just change the numbers to get it to fit whatever they might want to screw it onto...

But no.

There's no "thread" and no "hole" operations that actually accept parameters?  Only selection hard-coded size options from drop-down boxes shows up in all those lists...

So even though I can *measure* my threads (2.5mm on a 15mm outer-dia male nipple of 10mm height) it appears that Fusion360 has no support for actually making anything "thread-like" as parametric ?   ... 

Looks like it's first-principals time for me now... design something totally from scratch that uses paths and profiles instead ... sigh.  Since I happen to want to make it "concave/convex" (pipe-thread?) and think it's probably a good idea to go "complete" (design both the male and female at once) I'm pretty sure that's going to end up taking me all day, and I'm almost certain I won't be happy with the ends of the threads either, and I have no idea how I might manage to parametrically accomodate single/dual/triple/more concentric threads (each with tapered start/end) at all...

[MRWakefield]

I'm pleased you found it useful. The average person in the street  doesn't give screw threads a second thought but there has been a lot of thought and engineering gone into them.

Yes, unfortunately threads aren't parametric at this point in time but perhaps when AD add configurations to F360 this might become possible.

You asked about modelled threads in your original post: This checkbox determines whether the thread geometry is actually modelled (useful if you're exporting files for additive manufacturing etc.) or merely cosmetic (just applies a thread-like image). Modelled threads are more processor intensive so non-modelled is the norm if you're just producing 2D drawings for instance.

When you mention dual/triple concentric threads I think you're talking about multi-start threads. These can be created by circular patterning a single thread.

Good luck!! 😉

Hope this helps.

[TheCADWhisperer]

---

@TheCADWhisperer wrote:

Attach your new picture here.

---

Or not.

[HughesTooling]

@OceanHydroAU  Attached is a template I made as a starting point for custom form threads that might give you a starting point. Editing the pitch, OD and amount of turns seems pretty robust, if you edit the ThreadForm sketch to a different shape things will break but not too difficult to fix. The HelixAngle parameter is used to create a plane perpendicular to the helix so you get the correct form even if you have a course pitch.

[MRWakefield]

Great template @HughesTooling for odd threaded items. I'm not sure you need to incline the plane for the form profile though as I believe all the thread standards specify an axial plane for this. This would make the template simpler too.

[HughesTooling]

---

@MRWakefield wrote:

Great template @HughesTooling for odd threaded items. I'm not sure you need to incline the plane for the form profile though as I believe all the thread standards specify an axial plane for this. This would make the template simpler too.

---

You'll find the surfaces will be segmented if the plane is parallel with the axis. Like I have it you get nice clean surfaces you can fillet and tend to have less problems with combines etc.

Mark

[MRWakefield]

---

You'll find the surfaces will be segmented if the plane is parallel with the axis. Like I have it you get nice clean surfaces you can fillet and tend to have less problems with combines etc.

Mark

---

That's interesting, I wonder why that is? Anyway, it's still great and any error will be very minor so for most purposes it will be perfectly good enough.

Out of interest I took your template and edited it so that the thread form is on an axial plane and this is then projected onto the angled plane, so hopefully it's a bit closer to the standards. I've not done a lot of testing so I don't know whether it's any less robust than your original. Anyway, I've attached it in case it's of any interest.

[MRWakefield]

Ok, new version that adds support for multi-start threads.

[MRWakefield · Accepted solution]

... here's an updated version which seems a bit more robust when it comes to muti-start threads.

[MRWakefield]

I fear we may have veered slightly off the main topic so here's a scan of a nozzle thread I sectioned in case it helps.

 

[OceanHydroAU]

That nozzle thread... can you confirm that it actually *is* tapered (wider mouth, narrower back of the throat) ?  I just guessed it should be, since tube nipples do not seem to be...

[OceanHydroAU]

My Nozzle (printing now...):  https://a360.co/3sVyR4G

 

[MRWakefield]

I don't think it's supposed to be tapered. This was one that had split so is quite likely deformed. I just did it to see what the thread form was like. I found this item on Thingiverse but the thread form looked completely wrong to me, that's when I decided to have a rummage around and found that split one.

[TheCADWhisperer]

The threadform on the actual part appears to be a >> knuckle thread << (perhaps modified a bit) rather than v-form.