3D Printed Threads

I'm now achieving consistent results printing quite fine threads for camera adaptors. I have a T-mount thread 0.7mm pitch) here that measures 41.95mm OD and spins as freely into a t-mount adaptor as a commercial metal thread.

For metric form threads this is what I am doing:

Create a cylinder at least 0.04 smaller than the core diameter.

Create an equilateral (i.e. 60 degree included angle) to represent the cross-section of the thread, side length about 1.5 times the pitch.

Position this so that its outside point (crest of the thread) is 0.02mm less than the nominal radius i.e. the thread is 0.04mm undersize.

Use the spin tool to convert the triangle to generate a clockwise spiral of the right pitch. Make sure you have the highest possible number of segments or choose 'smooth' if you can.

To create an internal thread, create an external thread as above then increase the diameter to 0.04 to 0.06mm oversize. Do not alter the pitch. Subtract this from the outside of a hole that is about 'tapping size' for the thread.

Layers of 0.1mm give me a good result with a 0.7mm pitch thread.

Notes:

Obviously the printer needs to be well set up in X and Y axes with no slop.

Print dead on diameter using an extra outer layer, then tidy up with a tap or die. This works really well with coarser threads such as M6.

If threads curl upwards (i.e. look flat on one side) try using thinner layers.

To make threads looser or tighter change the diameter in small steps, say 0.04mm, never change the pitch or the thread angle.

That's an impressive demonstraion of precision 3D technolgy in the home isn't it! Very interesting, I want one...

I got one...

You won't go far wrong with the one Neil has. See Factory 3D Kit Build for more info.

John

Edited By Journeyman on 21/08/2017 11:43:37

If you print both sides of the mating thread it should work regardless.

I know of a gent who 3d printed a fountain pen, including all threads in it, and got the 'Shapeways' bureau to print it. They did a very nice job of it.

Regards,

Richard.

Will they make anything sufficiently durable for lead screw nuts, yet?

That would end the numerous problems of those that require left handed theads of some obscure size?

There are a number of quite exotic plastics available now as filament for 3D printers. Igus for example do *** this one*** which is supposed to be very wear resistant, similar to their Drylin bearings. They (Igus) say:-

"Our tribologically optimised filaments are up to 50 times more abrasion-resistant than standard 3D print materials. This results in higher levels of effectiveness. So you can produce prototypes and small series more quickly and cost effectively. Printed components can also quickly replace wear parts and reduce downtime periods."

Sounds good but I don't understand any of the technical info they supply on their website!

John

They do bronze which is actually printed in wax and then investment cast

JB,

Would that be sufficiently close tolerance for a lead screw nut?

Just make sure you catch them on a good day, they're running at full capacity most of the time!

Michael W

It occurs to me that anyone whose view of 3D printing is based on earlier experience of home-built 3D printers is likely to think I am talking rubbish, or that my 'thread; is just a loose engagement with the surface pattern on the print or at the very least feel that I am exaggerating. I ought to post a picture of the 0.7mm pitch thread to show what the thread form looks like. So here is one:

What is the material used and your nozzle size please.

.

That looks very encouraging, Neil

... it would be interesting to section a piece, and photograph in detail.

MichaelG.

Rod, it's a PLA, 0.4mm nozzle.

Michael, I have some experiments in mind, including a pull-out strength tests on threads.

Neil

Would you feel happy with, say, camera adaptors made using this system Neil

Here we go, this was a test piece where the bayonet mount was too tight. Slightly magnified - the threaded portion is about 4.4mm high. Not perfect, but being sectioned with a razor saw didn't do it any great favours.

It may not be obvious at first sight but it is easy enough for the software to make the filament size smaller than the nozzle size. If the extruder is travelling faster than the filament is emerging through the bore of the nozzle, it will be drawn out and end up smaller diameter. Which is how Neil was able to set the layer thickness at 0.1mm with a 0.4mm nozzle. Obviously, if you halve the layer thickness it will take about twice as long to print.

Also worth noting that you can control the density of the internal volumes of parts, resulting in either solid internals or honeycombed cavities. For strength you might increase the fill density but for speed and weight / cost reduction you'd want honeycombing. I expect the distortion (during cooling) could be a problem with completely solid parts - but my experience there was very limited so I couldn't pretend to offer any meaningful advice.

Murray

Sorry Murray, that's not quite right.

The nozzle diameter sets the 'default width' of the extrusion, typically the machine extrudes the amount of material required to extrude a rectangular cross section with the bead width matching the nozzle diameter and the vertical size equalling the distance between layers. The thickness of the bead is not strongly related to nozzle diameter and accuracy is achieved by the 'smearing' action of the nozzle tip. This is why I can print 0.02mm - 1/20 of the nozzle diameter.

In practice the extruded filament has (roughly) the from of a rounded-ended lozenge, like a running track in cross section. The fineness of surface detail in the horizontal is limited by the nozzle diameter. The horizontal resolution is quite fine and should not be confused with the smallest feature than can be printed, the limitations are in principle the same as those on milling narrow slots and fine features.

By slight under-extrusion the software will attempt to print widths down to about 0.5 x nozzle diameter. This is not always successful and it is best to make thin sections a multiple of the nozzle diameter in width, or use a finer nozzle.

This example of a thread prints well, because the horizontal shape is essentially just a series of slightly displaced circles with no sharp corners.

There is a great deal of 'tweaking' possible in terms of extrusion parameters. The latest version of Cura has a facility to allow for the slight 'oversize' that allows for the edges of each bead being rounded, not square. This subtracts a small, consistent amount from each layer.

Neil

<edit random smiley removed>

Edited By Neil Wyatt on 25/08/2017 11:05:10

.

Thanks for showing that, Neil ... I will send you a P.M.

MichaelG.

Nothing to do with printing threads but an interesting post over on MEM regarding costs and accuracy of the various types of metal printing via the online bureaux.

True enough Neil, although the law of conservation of mass presumably still applies - which suggests that if you are printing a layer thickness that is a fraction of the bore diameter (like a 20th), the bead is being drawn out, even if it is finally flatted into a pancake section. Presumably the bead is not being flattened out to 20 times the bore (8mm!)? I suspect the CSA of the extruded bead is somewhat less than that of the nozzle.

Murray