24 Hour Torture Test of 3D Printed Bearings
Вставка
- Опубліковано 29 жов 2022
- I am designing a brushless robot dog and needed a stupid amount of bearings, so to save some money I decided to try and make my own. Nothing crazy as it’s a simple model, but I made sure to test it for 24 hours and with 25lbs of hanging load to make sure it would be able to handle the forces I would be putting them through. Figured since they held up well, I would release the files in a bunch of different sizes in case other people could get some use out of it too.
BB's Link:
www.amazon.com/dp/B000HKKY7C/...
Printables Link:
www.printables.com/model/2632... - Наука та технологія
I wanted to correct a very obvious mistake I made in the video! I was recorded this after a long day of school and clearly my brain was gone.
the shaft was rotating at 11rpm which is NOT 110 rev/min like I said.
It is 66rev/min which translates to 95,040 revolutions in total over 24 hrs and 47,520 revolutions in total over the 12 hour span.
It's not a big problem. You can add a text overlay via youtube editor on place, where you said wrong numbers. I saw, many do that on their mistakes.
@@LichtTempler I tried to do that through UA-cam editor, but it seems they removed the annotation feature!
@@LichtTempler unfortunately they removed this feature some time ago
@@Dangineering you could pin this comment so more people see it
@@l0k048 Oh! thanks for bringing this to my attention. I thought I had pinned it when I first commented, but clearly something didn't work.
Nice work, I did a similar test, but at higher speeds and for less time. I’m always amazed by how well they hold up
Yeah I fooled around with them at several thousand rpm and like most bearings they certainly got hot after a couple minutes, but they are perfect for high load/low rpm applications or short bursts of high rpm during prototyping.
Did you lubricate the bearings before running them? If not, they hold up quite well! I'd imagine a bit of lube would go a long way
Yes, I lubed the bearings once with some silicone grease and then did not add any more during or after testing. For the amount of weight and duration of their use I was also impressed with their durability
@@Dangineering you should compare the results to a lubed printed bushing, i feel like the bearings might actually hurt because they point load the race
Dry lube like graphite would probably work well
Absolute fantastic work these bearings look great. I'm excited to see where this goes. (great Thumbnail btw)
These are super awesome. THank you so much for designing these and making them available. :D
This is beautiful work.
Been waiting for a video like this! Thank you!
These are awesome! Cost is definitely a factor when doing hobby projects, and this makes more projects more accessible for people. Thanks for sharing. Now what to build?
That was exactly the goal! Regardless of the project you end of building I hope these help
Very impressive what you did there!
Great video!
~66 RPM btw. Nicely done, and well tested. Bravo.
Was looking for this comment. Not that it really matters, but yeah. 66 rpm and arround 95k total revolutions
Your videos are super high quality, surprised to see only a triple digit subscriber count.
Thanks! I am fairly new to uploading, so I chalk it up to that.
@@Dangineering Yeah, you're definitely getting good engagement. I'm sure it'll pick up soon.
@@JimboJuice oh it will !
Very cool! Looking forward to the robot dog. Love me some robot dogs. Subscribed!
Nice job!
Nice design 👍
I would be very curious what dimension is distorting under load.
One idea I want to try, (or see tried) is to take advantage of thermoplastic creep to force mechanical dimensions that loose their true initial shape, back into optimal parameters. It would be kind of like a "self sharpening chainsaw blade attachment" but for plastic, either adding or removing plastic at specific points with a constant spring loaded force.
I'd consider using roller bearing design for the triple in future if can get ahold of the rollers they might held lower wear as the weight distribution is significantly better than a ball
Do you have any suggestions for what might make decent rollers?
Have you considered making the bearings out of the more exotic high strength filament?
Nice job with the metal bearings. I’ve made ball-less bearings for low load applications like the roller for my filament holder. I wonder how they would hold up on this kinda of testing.
I tried to put the same 25lb weight on the 100% 3D printed bearings that I found online, but they didn't rotate at all under that weight so I decided to not film it.
nice i like this idea a lot
Creative video, thanks for sharing :)
not just cool, rly cool. Attempted something like this long ago with plastic bb's so it was a chemical resistant being, didn't hold up to the speeds i wanted but i might revisit with your design!
That's why I stayed away from plastic BB's I knew that I needed them to take some higher speeds, forces, and temps so I opted for steel BB's.
Great video and testing! For the record, though, 11 rotations every 10 seconds is 66 RPM (not 110 RPM) so that's probably going to throw off your total rotations pretty significantly.
HA oops! Massive mistake right there. Thanks for the correction! I made the video late at night after classes and obviously my brain was pretty toasted by then. Thanks for pointing that out
@@Dangineering understandable :)
i wonder ifa carbon fiber or a glass reinforced nylon filament would be substantially better or not
Do these BB's have a 'flat'? (BB's used to). Only thing I can add to this EXCELLENT idea is you can get a bag of bearing balls which are mirror steel on AliExpress for really cheap. Totally going to try this.
You can have Chinese bearing very cheap provided that you design with bearing dimensions widely use. They will be better than any pla bearing, but pla bearing are very interesting for prototype and validation of concept .
Real bearings are and will always be better performing, but for the really large internal diameters that I needed it was not financially viable to use them. However for prototyping and simple hobby projects that will never operate in extreme conditions or for continuous operation these are a great way to save a buck.
@@Dangineering I agree I give you a tip and you give me one
What parameter do you use to have a smooth track for the ball ?
for my tip type
"axk official store" aliexpress ball bearing
-
you will have a 6810 2RS 50x65x7mm for less than 4$
It is not japanese quality of course but if you respect the load it will work
Really nice job. Those bearings are huge.
Apologies if you said it in the video and I just missed it, but how tough is it getting the bearings into the center race on the triple-bearing print?
It's not that tough honestly. You just put the bearings in on one side and then work your ways towards the other. It can take a little bit of patience. I will post a quick demonstration video soon sine a bunch of people are curious on how to assemble to triple race or larger bearings
Nice! If you make a pair of angular contact bearings, that would allow you to simply squeeze out any play that might develop over time.
Interesting Idea! I will give this a try in the future and until then I would encourage people who needed minimal play to simple make multi-layered bearings as they have surprisingly low play .
@@Dangineering will you prototype angular contact bearings as well? They are the best for robotics applications
@@gedr7664 Yeah, I will work on a design over my winter break and if it works well release it in a similar manner to these deep groove ball bearings.
I have a small PLA part for my home that kept breaking, and i just kept printing more. I find pla can be quite brittle and thin parts like this can be especially fragile, but I guess depends on the exact plastic and perhaps the degree of crystallization. But I tried the fancy 3d870 pla resin, annealing, nothing seemed to work. I switched the part to a PHA resin from regen and It hasn't cracked since.
Personally I have had a really good experience with PLA when I increase the width of the walls. I recently switched to PLA + or PLA Pro (same thing) and have not had any of my prototype parts break on me.
I'd like to see the method for packing the multi-layer bearings.
Working on getting one up since I have had a lot of requests on triple race bearing assembly.
Very interesting! A few thoughts on wear:
- lubrication (obviously)
- Nylon (you need a nylon-capable printer obviously, which I don’t yet myself)
- I suspect that much of the wear is just the high parts of the layer lines wearing away. I’d think that the wear rate would decrease once the high points are worn down. I wonder how it would work to print with slightly over-tight clearance and then sand the races lightly to give a smoother surface?
- someone else mentioned electroplating, but I wonder if some UV-curing resins might be harder than the PLA? I’m thinking in a similar direction to the last note: just fill in the low points in the layer lines a little bit to provide more contact surface area.
- I wonder if the smoother surfaces of actual ball bearings would reduce wear? BBs are certainly cheap and pretty smooth, but extremely smooth vs just very smooth might make a difference.
(None if this would be an issue for your robot dog use case; you’re not going to be running the joints 24/7 for months on end :-)
Great, great work!
My impression is that UV resin might be more brittle than PLA, but Nylon should be very hard wearing. Maybe a design with cylindrical roller bearings would wear less than ball bearings, but that probably introduces other problems.
@@DisorderedArray Ah, good point about the brittleness of the resin. I think you’re right, UV resins are known for their brittleness, and the high point-contact forces of the ball bearings would almost certainly cause cracking and very short life. Thanks for sharing the comment!
@@DisorderedArray BTW, I just came across this design the other day, using wire for the bearing races. Still very low cost, but they should be much more rugged: ua-cam.com/video/xbGtqpgfdRM/v-deo.html
Very interesting thoughts! I will definetly try and print it with slightly too tight of clearance and then sand away a little bit to test the high layer line theory. I would like to try electroplating, but practically that is not a real option for most people (so even if it is a solution, its not a viable one)
@@DEtchells interesting, I guess with these larger 3d printed bearings one has to match the design choice to the expected wear profile.
Are aerodynamic air bearings possible with PLA for high rpm?
Does plastic wear less the more surface area is contacted or would that just create more thermal issues?
I am really impressed with the result, and I am greatly interested in plastic made bearings, for rust proof bearing, using stainless steel BBs, for creating my own ECM lathe, (I will just have to find a way to get current into the immerged chuck, without any contact with the highly salty water solution as I would get extremely quick corrosion, maybe thanks to a usual steel bearing encased in an oil filled cavity, I will see)
but I have a little question : how do you fill the multi-layer bearing of it's ball bearings : is the balls separator in several pieces and each level gets filled mid print ? Is it filled post-print going layer by layer with a splitted ball separator ?
Thanks a lot for the video, it gives a lot of ideas
Exactly. For the multilayered bearing all you do is print one casing and three retainers. After they are done printing put the BB's in one of the most external races and then go layer by layer till you reach the outside bearing race and boom you are done. If you wanted to wait a little bit I am going to post a quick little assembly video for multi layered bearings because it seems a fair amount of people are asking questions about it. I will be curious to see if these could do well in a ECM Lathe environment
Is this going to become a song I’m waiting
If you want some more axial load capabilities and still be single groove, you could design something like a magneto, angular contact, or deep groove. im sure you already know of these just thought I would mention it
Thanks for the suggestion, I will look into making an angular contact version that could maybe offer more axial performance in the futue
This video showed up on my feed and I just clicked on it while not really thinking about it. I had no interest in ball bearings whatsoever and am not building anything. Now I am 12 minutes in and must say: I kind of got interested. :)
Haha, that's the UA-cam algorithm for you. I have done that same exact thing so many times. Glad you enjoyed it!
5:00, “11 revolutions every 10 seconds”, that isn’t 110 rpm, there are only 60 seconds in a minute, it is 66 rpm.
“Thin PLA 3D Printed Bearings”, then prints the strongest out of PETG.
I acknowledged the mistake in my pinned comment, made the video late at night and clearly was tired.
Yeah I simply ran out of PLA filament and wanted to still show that you could print multi layered ones.
@@Dangineering didn’t see that when I commented, also it still isn’t quite right, you said in the pinned comment that the shaft was rotating at 11 rpm which is 66 rev/min. It is a great project though.
when you print multiple layers of the same print, do you need to print additional bearing cages as well?
4012 lane bearing needs to be a thing
Very nice bearings. The shape looks similar to a ceramic bearing I’ve got in the workshop. Do you print them 100% fill? Subscribed, I love this kind of thing :) ps does heat become a problem?
Glad you enjoyed it! Yes, I printed them at 100% infill which I achieved just be increasing the amount of perimeters. Heat didn't become a problem for me at these speeds, but I have taken one (without lubrication) and had it spin at around 8000 rpm for about two minutes. At that speed I could feel it slowly start to heat up to the point where at the end of the 2 minutes the PLA was a little soft and flexible. if you need higher speeds I would suggest to use nylon, lube, and if you need constant operation for long periods of time to have some sort of cooling.
Did you lube your 24h bearings?
Amazing! Could you please tell us where did you get the balls and what are they made out of?
Here is a link to the exact BB's that I used: www.amazon.com/dp/B000HKKY7C/ref=twister_B081ZSPJ9Z?_encoding=UTF8&psc=1
have you tried using filament cut to length as roller bearings?
I have not tried that, because I fear that the filament might get a little too hot and then start to get too flexible, lose its shape, and then cause the bearing to fail.
on voit des encoches de remplissage pourquoi merci
thats metal ball in those bearing? i put the printer in pause at half way then put the metal ballz by hand and continue the printer?
I actually made a video about this: ua-cam.com/video/5FTnaDngUjw/v-deo.html
nice
How do you double or triple these bearings? I apologize and advance for this stupid question.
just import the STL or STEP file into your soild modeling program of choice (OnShape, Fusion 360, Solidworks, etc) and then create an assembly by having multiple of the casings concentric and then make the sife face of it coicindient with another one. Then you have what you need and you can make them as large or small as you need.
You probably thought of this already, maybe ruled it out because of difficulty getting the pins, but could needle style bearings with pins instead of balls wear better on the plastic?
Seems like the spread out contact point would work better for plastic.
BTY no engineering experience
Did you lubricate them before running them?
Yes, I lubed the bearings once with some silicone grease and then did not add any more during or after testing.
I want to build some larger bearings. Where do you buy the ball bearings?
Any Steel bbs that are 4.5mm will work perfectly.
Here is exactly what I used:
www.amazon.com/Crosman-Copperhead-BBs-EZ-Pour-Pistols/dp/B081ZSPJ9Z
@@Dangineering Thank you. I like your videos and subscribed. I just ordered those balls. 70% off $2.98. pretty good deal!
@@FilmFactry thanks! Got some really interesting videos coming up soon! Hope you enjoy them! Also that’s a great deal! Make every bearing you make only cost a couple cents.
@@Dangineering Got my ball bearings today. It would have cost more than $3 just for postage. Now to print a bearing to test. Thanks.
@@FilmFactry Congrats, that sale was super nice! I would love to see how they turn out once your done! Also the larger you print the more flexible they become unless you bulk it up by making it multilayered or increasing the wall width.
How did you assemble these bearings?
Actually made a video for this: ua-cam.com/video/5FTnaDngUjw/v-deo.html
It looks like the balls could be spaced closer together. This will spread the load out better and should decrease the wear quite a bit as they ball do not need to sink into the race that much.
I was not able to generate enough load to stop them rotating as is. So I figured I would get more benefit from having slightly thicker plastic sections between the balls to decrease the chance of small bits breaking than I would from increasing the potential load capabilities.
@@Dangineering He is not talking about increasing load capabilities, he is talking about reduced wear and i think he is right. Mechanical engineer here. I think sections between the balls can be a little bit smaller without any problems. You should test it. :)
@@drister007 Yeah I understand what he wanted, but got lazy and didn't explain all the horns of my argument. It would likely reduce the wear, and I might test it in the future, but I will probably take me a while to get the time.
If I get some time over winter break I will make some variations of the retainer ring with higher density.
@@Dangineering Cool :)
cool
emmett made a OpenSCAD file for his gearbearing. you can actually get away with very big/thin bearings. Try playing around with it a bit maybe.
edit. i really like your approach tho, the tolerances of the bearings seem really good. roundness seems to could be improved a tad but still very nice :D
also PETG is quite durable for such applications and lasts in mineral oil for ages. nylon (PA12 or PA6) would be okay too but PETG is overall a lot less hassle to get tuned in properly.
And for the sizes of bearing you would need. they are expensive as heck. sadly.
peace o/
I will give that a look! Yeah nylon will be the best option for ultimate durability, but man is it a hassle to get good prints with it sometimes. OH yeah the bearings I needed were just prohibitively expensive.
@@Dangineering From my own experience (some gear sets that now run a bread slicing machine) and few different gear and cylindrical (more surface area) bearings IMO if you have say a "light load" application PETG will usually hold up WAY better then PLA and not that much worse then nylon. usually my overall go to material nowadays. What printer are you running if i may ask ?
I see the casings and retainers in the downloads but what about the balls?
The balls are made out of steel and can’t be 3D printed. However here is the link to the exact ones that I used:www.amazon.com/Crosman-Copperhead-BBs-EZ-Pour-Pistols/dp/B081ZSPJ9Z
Would be interesting to use them on a skateboard
That would be quite the ultimate test of their durability! Would love to see how long they hold up
Put a magnet on it
This was great work(!) I just came across a design yesterday that uses wires for the ball races that you might also be interested in. They add another step (you need to ring-roll the wire), but still extremely low cost and they should be much more rugged than having the bearings run directly against the plastic: ua-cam.com/video/xbGtqpgfdRM/v-deo.html
(The creator says he’ll share the F360 files with people if they ask. He’s apparently not posting them himself because they’re highly parameterized and quite complex, and I guess he doesn’t want to be swamped with random questions.)
The straight plastic bearings you designed are almost certainly fine for your robot-dog use case, but I thought you might find the wire-race approach interesting 👍😁
I had not seen those before, thank you for introducing it to me! They are certainly way more work to assemble and seem to use more hardware than mine (in that you have to screw it together), but ultimately I would guess that if you need slightly higher performance then his might be a better choice.
Those will likely last longer, and tolerate higher speeds and loads, but even my own design exceed my personal force requirements, so I will probably keep the design as it stands unless other issue pop up in the future.
Where can i buy those bearings like each single one
You can 3D print them for next to nothing if you have a 3D printer and some BB's. Check the description!
Oki thanks
I don't have a 3d printer though
Which ones would fit on a 4 1/2 inch cylender and a 5 5/8 inch cylinder and on a 6 to 7 inch cylinder
Im trying to use bearings for a metal suit I'm making
Will a stack of triples (like 3 or 4 triples) hold up to a lathe headstock if lubed well?
Not a chance, PLA simply isn't hard enough. They would wear very quickly under axial load, and they'd also provide a fraction of the rigidity of steel bearings.
It probably won't for long term use, but if you make a stack of 3-4 with much thicker walls than you might be able to squeeze out a little bit of use, but just a couple minutes if anything. (If you try this I would encourage you to use nylon or something that can tolerate higher heat). No guarantees, but I would love to see the results.
@@Dangineering Nylon wouldn't be much of a problem. If I try this I will definitely be hard nickel coating a few thou thick. Spray a thin layer of copper conducting paint and then electroplate a "thick" layer of nickel chloride or the like (can't remember exactly which is hard nickel vs decorative nickel)
@@natereinhold6180 That would be super interesting! I would actually be down to give that method a shot should it work for you. Have you done this coating process before on other 3D printed parts?
@@Dangineering Electroplating 3D prints has been around for a few years. There are videos showing the results here on UA-cam.
you need a dialgauge and a press :D
That would have been very helpful! but unfortunately not in the budget..
@@Dangineering dial indicator isnt that expensive. You dont really need the mitutoyo. And you can make a hacked press with weights and a lever. Not expensive at all. You can do it!
I think cooling is very important
Yeah, at very high RPMs (8000rpm) they start to get hot in about a minute and start to lose integrity during the second minute unless I started to cool them. However at those speeds I would advise switching to a real bearing as these are more for functional prototyping
@@Dangineering where would you need 8000 rpm? already 2000 is enough revolutions like a lathe
@@martinTintin476 I am using a relatively high power brushless motor in my combat robot and was using it as a means to informally test the bearings a little bit. I don't need that speed for any of my projects I just wanted to share my experience.
11revs 10s
110revs 60s
🙄
when I realized that mistake I made the same face.
Petg has a higher melting point.
Your math is off. 11 revolutions per 10 seconds doesn't equal 110 revolutions per minute as there's only 60 seconds in one minute.
This is not a problem you can solve with a 3D printed part in plastic. Search for "Frictionless Bearings - Technical Secrets Explained!" here on YT.
24 hours? might es well be 5 minute test when it comes to bearings. Go skate with it and report back 😁
Never claimed I could solve anything?
These bearings are not meant as a 1:1 replacement. They are meant to be used as prototyping tool so you don’t have to waste money on expensive large diameter bearings that won’t be used in the next iteration of the design.
I personally would not skate on these as I don’t think they would last, but I do think it would be interesting to see the results.
I’m not trying to design around real bearings with PLA. I’m trying to give myself tools to allow myself to design with parts far outside my price range
Great video!