3D Printed Ball Nut for the Insane

[tom osselton]

Looks like you are one short!   

[PK]

You got 2 sizes?
Every second ball in a ball nut is a metric poofteenth smaller. This allows it to act as an idler between the load bearing balls.

[Joules]

   Oh no.....

I'll widen the raceway every 3mm, that way each one can be alternately idle.

[Joules]

    PK, I had to look that up as I thought you were pulling my plonker....

Well friction will take care of itself as the nut heats up.

[awemawson]

Resurrecting an old post here, but where are the pictures of the final result Joules 

[Joules]

Revisit time for this mad project as I now have the 0.3mm nozzle.  First things first, find out what the real hole sizes are that I need for these bearings.  It's one thing doing a design in CAD but another doing the prints, so a ball gauge is needed with holes in Z and Y from 2-2.5 in 0.1mm steps.  Why Z and Y, because printers don't print true, they normally do a good round hole in Z, but you end up with a droopy ellipse in X and Y.  This will impact on the raceways within the nut.  As can be seen in the pictures, balls in the Z pass through at 2.4mm and pass at 2.3mm in the Y.  I was expecting the opposite result, however that sets my minimum hole size at 2.4mm.  That means the nut needs redesigning to take this into account.  I will also need to make some test nuts that just have the spiral raceway in to set them, it could be they have a different diameter or spacing from the holes depending on how the balls peen the printed surface.

I looked up ball options as per PK's original observation that I needed different size balls.  The nearest I can find are 2.45mm and 2.5mm, but in a printed nut you can't hold that close a tolerance so we go ahead with full contact balls of 2mm.

[Joules]

Well that was interesting.  I printed my first test nut and found I printed a mirror of what I needed.  Doh....   So I tested the nut for fit on the lead screw noting the threads running in opposite directions and marked the nut as bad.  Then I had a thought, well a single bearing should cycle through nut with no motion.  It did, and I could feel the quality of the race within the nut and it was good, with no tight spots.  I will cycle this a few times more to see how it beds in.  The nut was tight on the shaft so the next MIRRORED nut has 0.3mm extra clearance but same thread profile and position within the nut.  No play between the nut and the shaft with the bearing inside, looking good.

Doesn't mean a whole bunch of balls won't lock up in the correct nut, but we should know in another hour or so.

Well bugger me !!!   loaded 13 balls in the test nut and it seems to be working, I wasn't expecting this.

[Joules]

It works !!!  Loaded the balls and they are recirculating.  At the moment the ball nut is still a little rough, but the bearings are making the track smooth.  I have enough clearance using a 2.3mm track throughout.  No backlash in the nut, yet.

[S. Heslop]

That's pretty awesome! Interested to see how it holds up in testing.

[Joules]

After hours of running the ball nut along the shaft it has flattened out all the internal imperfections.  The nut isn't as free running as a metal nut, it was never going to be.  It does have zero backlash and should maintain that in low load applications as the nut contains around 70 balls, lightly greased so they slide against each other, but don't seem to slip in the nut or on the screw.   I have published a couple of STL files on the GrabCAD site listed earlier in the thread.  They are this new nut and the ball gauge.  Don't just expect to print this nut and it to work.  The gauge was used to set the raceway dimensions based on this printer and filament.  The nut is scaleable if you have the CAD skills to draw the raceway, I think my 2mm balls are too small as they sit in the lead screw and they should ride on the edges.  But hey, it works for me, don't underestimate the importance of using a 0.3mm nozzle with balls of this size, that blob size will catch you out.

[Joules]

Ahhhh, inspection time showed up a couple of flaws.  I had been stuffing balls in to reduce the gaps, still had about 1 balls worth of gap and I noticed the ball nut has started cracking.  Fair dues, it's not much thicker than 1mm at the thinest points between shell and outer, and it was designed rapidly last night as proof of concept.  I could print thicker shells as this is pretty thin, increase the raceway clearance a gnats or machine an alloy case to clamp the nut externally.  That last option isn't a bad idea as the nut becomes an insert that is easily replaceable.  Not quite an all printed solution though.  I have printed a second nut and taking extra care to prepare the internal raceway and scrape away as many fine blobs as I can find as I guess the pressure inside the nut must get pretty high as it's bedding in.

The cracked nut does feel smooth and still no backlash, time to pull it apart and inspect the raceway internally.

The nut internally is quite badly contaminated with PLA debris forced out of the raceway.  It was soft enough you didn't feel it, but I need to find a way of burnishing the race before assembly.   this nut is being cleaned up and I will carry on testing it.   The count was 74 balls in this nut, so thats about 150mm of raceway wrapped inside this tiny nut.  Need more balls for more nuts

[Joules]

Problem solved with nut No. 4

It now has a thicker body by an extra 1mm, I was also able to increase the raceway separation to 1.5mm, from 0.7mm.  The new raceway was increased to 2.4mm, like I said it should have been in the first place 

The temperature I was printing at was the high temp I used for printing the shells, this was causing the extra blobs I was having in the raceway.  The nut has had some running in on the lathe at various speeds to bed it in.  The photo below was after the running, and no fuzz from the PLA, so it looks like the raceway got burnished properly this time.   Printing more nuts for future use, and get some balls ordered.

[Joules]

I figured I would print a window to show the balls recirculating and give some scale to the nut.   This is the new revision 4 nut, whilst assembling the new window I fumbled the nut and lost 8 or so balls.  I was working over a tray, but most of what fell out only a few inches above, made a mad bounce for freedom.   The new nut has 84 balls in it, I am down to 5 spare as every time I open this, I loose a ball or two, or "8"

This has been a long time in the making, from those first drawings to not really believing this would work, to not an insane Ball Nut anymore.   I will carry on refining this design and try to find a worthy project to make use of them, or see if anyone else runs with the idea.

[awemawson]

Looks excellent Joules, do you have a real life application for them, or is it just the joy of achievement that floats your boat?

[Joules]

Design looking for a purpose.   I had an idea in mind but ball screws and commercial nuts put the project outside my play money.  I did still want to know if I could design and print a working version.   They might make a good alternative to belt drive or get a small desktop cnc working again.  I ordered some more balls, I should really print one and set it up with a thrust bearing then torque the ball nut to destruction.  I might also have come up with a serious use for IGUS filament.

[slowcoach]

Brilliant that Joules 

Rob