Thank you, bud!

Thanks, bud! 😊

Thanks again! The major concerns, after a while playing with these, are: 1. friability of the plastic (PLA is a no-no, petg is much more recommended) 2. the radial stability (a poor bearing installation will result in an easily ovalized bearing) 3. the inherent noise/vibration that is greater than on a precision ground steel unit Other than that, since the balls create their own raceways into the wires in time, a smaller concern relates to the need for checking / re-tightening the adjusting bolts after a while.

@@ErosNicolau That all makes sense. I’d think that friability and radial stability could less of a concern depending on how the bearings are supported by whatever they’re being built into. If they’re integrated directly into the mechanism, then yeah, you might need to go with PETG for everything. OTOH, PLA is hella strong and rigid if you just use enough of it. - Basically lots of perimeters; you want essentially solid plastic for a good ways around any stress points. You can use light infill to save on weight as long as the shell itself is very strong, although of course YMMV depending on the specifics of your application. Still, 1/4 - 3/8” (6-9mm) of PLA is pretty strong and 1/2” (12-13mm) can handle a *lot* of stress. I’ve made beefy connectors and braces out of PLA for a largish structure (~6x6x6 feet or 2x2x2 meters) made from 8020 extrusions and they’re very strong. Another factor with any plastic is creep under load. I think it was CNC Kitchen tested different plastics in that regard, and while I don’t recall all the results, I think PETG was one of the best. (Definitely check out his video for yourself though; my memory is pretty faulty :-/) PLA creeps quite a bit initially but settles down after a few days. I expected to see a good bit of creep under the bolt heads on my PLA 80/20 connectors, but there was surprisingly little despite me cranking down the bolts very tight. It was solid PLA about 6mm thick under the bolt heads though, and solid plastic for a good ways around them. I think the surrounding unstressed plastic may have constrained the part that was under the extreme compression and this prevented plastic flow from occurring; the compressed plastic basically didn’t have anywhere to go. It’s interesting about the balls essentially cold-working their own raceways over time. Thinking about it, you might not want to just let them wear-in in the actual application, as the races could end up asymmetric or ovalized from asymmetric forces acting on them. For best results, you’d probably want to lightly press them into a rigid jig of some sort and just crank up the pressure evenly with the adjusting bolts as you run them in over time. Anyway, this is a super-interesting design(!) I don’t have an immediate application, but am really intrigued by the possibilities of the approach.

@@DEtchells Wow it's so nice to see so much thought put into your reply! I'd like to answer some of your points: Indeed this partially 3d-printed bearing is more of a skeleton - it should rely on its host application too for extra strength - after all, that's how integrated bearings work. About plastic creep: I'm sure that's a factor too - pretty conveniently mitigated with the adjustable preload design ;) About the balls cold-working the wire races: the wire races sit into their channels a bit imperfectly (because 3d prints are not perfect) and the balls, after not so much effort, tend to quickly push the wires to their final positions in the grooves and to also create their own grooves into the wires. Which is perfect. I'm split there on your suggestion, as pressing the wires into their plastic grooves slightly deforms them and you end up with already work hadrened races which are then much tougher to reshape. Wearing them in in their plastic place instead has the advantage that you start from both more maleable steel AND single point contacts between ball and wire (so higher pressure available) About potential designs: I originally created this as a high ridigity slewing ring for a camera gimbal, but I can see this as a pretty good application for robotic arms where you potentially need both light weight, rigidity and large diameters. Esentially, I see this as fit for big diameter and slow speed applications.

Absolutely! It improves the smoothness and reduces the friction significantly. After toying with both bearings lightly for a while they are butter smooth now

Hey, bud, thank you for the suggestions! Absolutely, this can be improved upon based on the specific application. In this case, the application was simply to test the feasibility of the hybrid system and to showcase the process, so nothing fancy. This type of bearings are meant for slow speed and high load, so impact wear at the gaps on the balls is not much of an issue. Furthermore the wire gaps will always be better left small, as every engineer mind has suggested, but the less intuitive fact of the matter is the balls adjacent to a gap share so much of the load that the gap is basically inconsequential.

@@ErosNicolau WOW! Not just very good suggestions but also a perfect answer! I'll take all this into account when designing this for my own project...

@@radicalphil1871 Thanks, bud!

Oh and great progress! Thanks for posting these!

Thank you! Actually that's pretty imperfect and a source of friction. I will attempt a larger holes version tomorrow. The cages don't get scraped because the balls don't have sharp corners ;) and mostly because lubrication: the more the cage is conform to the shape of the ball, the greater the contact area but at the same time the greater the lubrication

Thanks, bud!! Ofcourse, this is not about a design, but about proving this can be done. Anybody can take the design further and add whatever helps. If you ad the vents in a nice relationship with the preloading bolts, then most probably there won't be any issues, structurally. About embedding the bearing into another unit - that's one of the main use-cases I see for it. I contacted a number of amazing people working on cycloidals - starting with Paul Gould - which in my opinion is pioneering the cycloidals the most - and a few others (I believe I left a comment for James Bruton too, but he's such a popular guy that it's easy to lose track of most comments) but I've yet to see anybody implement these into their robotics projects. I do believe it's only a matter of time though before we see such a steel wire-race bearing implementation in an amazing project.

Thanks for the tip on Paul Gould, I hadn’t been aware of him previously! Yes, I’d really like to see someone experiment with this concept for cycloidal bearings too, and agree that the use case where at bearings are integrated directly into the structures they’re supporting should be a popular one.

I made the whole thing parametric, so there's no STL, just Fusion360 files which are pretty involved. Drop me a line on FB Mess and I'll tell you more about it there.

Thanks, bud! To answer your questions: - I used zinced mild steel 2mm wire (it sells in huge rolls but also by the meter, over here). It's better that it's mild, because in time the balls press in their own grooves which improves both stiffness and smoothness - I ran the bearing on my Makita DHP484Z on the high speed (2000rpm) for 5 minutes. The more I run it, the smoother it gets (see point above). Note: every once in a while you need to re-check and re-tension the bearing bolts, since with the grooves on the wires being pressed in with time, the slop increases. But this phenomenon is on an inverse logarithmic scale: most of it happens in the very first minutes/hours of running, because there is less metal to be pressed in direct contact. Later, once the grooves are formed, they tend to stabilize because the ball-cage contact transforms from a point to a line. - This bearing is not for replacing traditional ones, but for specific, exotic applications, like large diameter units with a small cross-section (which usually cost A LOT), or for lightweight loads or where a lightweight unit is needed (say, for portability), or for places where you could really use a bearing integrated into the body of the unit.

A particular use-case for an exoscheleton can be seen here kzfaq.info/get/bejne/gbpjdaWBq9isZ4U.html

@@ErosNicolau That's a whole lot of recommendations! Cheers mate!

Thanks! :)