This project is THE reason I purchased a 3D printer...
My parents managed to save a wonderful plastic toy from my childhood that I would like my daughter to be able to enjoy, and as far I can tell it does not exist to buy and the original plastic is getting brittle and cracking. Kiddicraft is the company that produced it (apparently the actual inventor of Lego) in England. I have found very little information about the original product, comment if you had one, and if anyone has the story of the original design, share it! Let's put the historical record here.
After a couple months of learning, printing and hacking around in OpenSCAD, the result is a fairly accurate reproduction with 2 versions of some parts, one reproduces the moulded plastic the way the original toy is made, and the other is an FDM 3D printer friendly version that is just as good, if not better than the originals for strength and safety. Another change I made is the nut parts have threads all the way through, whereas the original parts have the barest whiff of a thread right in the middle, mostly because of the limitations of the way they were manufactured.
Take care printing the threaded rod part, have a look at the Print Settings section for some slicer ideas. I have added a shorter rod (that still fits all 3 nuts) for those having trouble getting the full size one to print and also activated the customizer with the relevant OpenSCAD file so you can create a custom rod of any length.
Big thanks to the creators of the Poor man's openscad screw library and Nut Job code that got me going on this project.
Anycubic Kossel Linear Plus
0.1mm-0.3mm x 0.3mm-0.5mm
Nothing special needed (0%-20% infill) to print the solid nuts other than brims which can help keep curved corners from lifting/warping if that happens. I accidentally printed a curvy nut with 0% infill and 1.2mm top/bottom (0.2mm layers) and 1.5mm shell (0.5mm nozzle) and it turned out great! As long as your printer has good part cooling for bridging, it should be good and solid without being any heavier than necessary. I think the original style hollowed parts can be printed in 2 parts and glued together for full quality, but I didn't see a good reason to print them like that myself. To do it (until I post a split part) make 2 copies in your slicer, flip one over, then cut off half the height (z) so that only the top part of each half prints. The cone nut can be printed with 0% infill and NO solid bottom to emulate the original form and weight, but I recommend solid bottom and top since it's safer and much easier to keep sanitary - but not too much heavier. The outside shape is quite complex, so it's a good idea to have the printer (at least mine) slow to around 10mm/s for the OUTSIDE shell since the jerking around of going any faster tends to encourage blobs oozing from the nozzle and forming small lumps on the skin.
Give the threaded rod settings in the slicer you know will produce a strong part on your printer, like 100% infill, and slightly higher temperature for better layer bonding and slow as frig. I found a raft and z-hop helped keep it stuck to the build plate. I also used the Cura tweak at z plugin to get it to slow down by 50% when it hits a certain height before it starts to get the weeble-wobbles.
try the NEW solution found to the problems caused by printing tall cylindrical objects: spiral infill... sort of... Basically turning off ALL normal infill and using the wall thickness to produce a solid or hollow rod. The problem I was seeing was partly related to the motion of the print head (delta) when doing infill since it always moves back and forth (wobble inducing) whereas the wall or shell follows the shape of the outer contour, so it moves in circles and only ever moves back and forth when retracting and travelling, therefore above the print by the z-hop amount. Making those changes allows the rod to stick to the build plate without needing the slight flex and bond provided by the raft. Speeds can be greatly increased as well since the forces involved won't break it free as easily. It may also mean being able to print a longer than original version, but it may not be strong enough to be considered kid-safe since much more leverage can be applied to a longer toy by a child stepping on it.
I have tested printing it completely solid spiral, and also as a tubular object... A few more settings in the slicer need adjusting for an infill-free hollow version, so I will post a Cura screenshot in Post-Printing (below) of what to look for. Obviously hollow is not as strong as completely solid but seems stronger than 50% infill and lighter, but also allows for a MUCH quicker print (30-40mm/s vs 5-10mm/s), take a look at the video clips if you are curious.
Uh oh! Forgot to turn infill back on after printing a hollow threaded rod... Happy accident.
For some reason I didn't think it would produce a sturdy feeling hollow part, but it worked wonderfully and feels nearly as light as the original part.
Clear and 'solid' green, both 0% infill.
Clear green looks darker in some lighting conditions because there are no white particles to reflect the light.
Shell thickness is the infill, so make it much thicker than normal, no solid top or bottom and 0% infill, should result in a nice strong tube.
Short clip of the full size threaded rod finishing, printed at 40mm/s !!! Layer height was 0.1mm so it took 7 hours but it looks amazing!
Short clip of a hollow rod with no infill printing faster since almost all the printing motion is circular..
Short clip of a solid rod with no infill printing: I set the shell or wall thickness higher than the radius.
Clip of a threaded rod printing at 10mm/s (with 50% feed rate reduction for height) with normal infill. Must go slow to prevent the back and forth motion from snapping it off the build plate.
The original plastic is now easy to break, and was glued to keep it intact, hence this project.
I tried using the Nut Job customizer to reproduce the shapes but I couldn't make it generate two non-threaded ends, so I went on a 2 month long quest learning OpenSCAD, also learned how to measure threads properly, one is not enough!
Also, I couldn't get it printing the rod, such a high thin object without it snapping it off the build plate... I really had to slow things down and use z-hop.
Close but not quite! Print, measure and adjust.... render, export, print again.
Speaking of rendering... 9 hours? Time for more RAM?
Countless hours spent coding it perfectly hollow and rounded, only to discover I could not print it properly as a single part.
Hollow: solid top & bottom, 0% infill on the left, original part in the middle and no solid bottom 0% infill on the right.
Some rejects: learning from mistakes... printed the rod way too low temp to avoid the hot thin edges curling up, but lost layer adhesion. Print it for strength, not accuracy - there's 1.5mm slack because it's a toy that's meant to be easy to spin.
The problem with air-gap rafts and circles...
13 minute collage of mostly video, documenting some of the 3D printing journey I went on, at least when I remembered to shoot video. Includes a slow-mo of z-hop.
Updates and Proposed Changes
This section will be to catalogue changes and announce anything I'm working on before it gets updated. Please comment if you have an opinion on the proposed update.
Below you will find my current adjustments I'm testing to get the shape slightly more accurate. The render time is a little insane (over 9 hours on my i7 laptop) with all the minkowski rounded edges, so the process is slow, but then again so is printing.
I think I've made enough adjustments to warrant a release, but to let people decide for themselves, I'm leaving the first version so you can download them both and compare.
Print of the first STL on the left, original nut on the right.
First version STL on the left, new version on the right.
Same as above but an un-distorted by perspective (Orthogonal) view so you can really see the shape difference.