Replacing Drivetrain Parts with 3D Printed Carbon Fiber Nylon - FOR SCIENCE!
HTML-код
- Опубликовано: 4 окт 2024
- Will it work to replace a differential ring gear with 3D printed plastic? how soon will it break? lets find out!
CF Nylon (NylonX) : goo.gl/a6zeQd
Special Nozzle for NylonX: goo.gl/WTNhxq
Check out my Patreon Page! / geardownforwhat
Buy a Ratcheting CVT T-Shirt: a.co/gjs5CTY
Follow me on Twitter
/ gdfw17
I post most of my designs on thingiverse
www.thingivers...
Would you help translate my videos? : www.youtube.com...
Links to products in this description may be affiliate links. These links do not add any additional cost to your purchase, but the channel gets a small kickback, which helps us a bunch! Thank you!
Hey! You guys talked me into revisiting this, I'm currently reprinting it! Follow me on Twitter for updates: twitter.com/GDFW17 and don't forget to ring that bell to get a notification when the video comes out!
Man you should print one solid piece and then tap the threads OR if you really want solid threads then check out V-coil, it's basically a tool that allows you to repair threads if you'd see it you'd understand how it works, I use it all the time in the workshop and those V-coil coil threads are really strong and I think they would really strenghten that part, you'd just put the coil threads into the plastic. Cheers!
Could always do it the same way, but just use a heat gun to reheat the cooled down material.
Or what if you made it as a 2 piece mold, like what it turned itself into when it broke, but with thinner nuts, and less thickness on the mating face piece, then a bit more on the toothed section, and drill spots to put some steel guide pins. Or print them onto one of the parts, but any sort of pin setup would hold the 2 pieces together better.
Another way is to insert the nuts while printing. I would restructure the program to print a complex layering sequence to allow for the insertion of the nuts while maintaining structural rigidty whilst in perpetual motion.
The best and probably easiest way to do this would be to slice the model in half, then add protrusions to one half and divets that line up with the protrusions to the other half leaving holes half the thickness of the nuts on both halves so that they also act as pegs to hold the layers together.
If you send me an stl file, I can fix the model for you. :P
Gear Down For What? Post this on thingiverse and I'll tinkercad it to perfection for you.
So... why the need to encase the nuts? The NylonX layer strength is always going to be better than JB Weld. Why not make pockets to slide the nuts in? It's possible I'm missing something, but, I know one certain fact..
THIS WAS AWESOME. You need to do it again. Seriously.
I will try again if you collaborate with me and make a version out of polycarbonate!
I tried soooooo hard but failed miserably sooooo many times and gave up!
Gear Down For What? Polycarb? Dude, great idea!
Using pockets would be great, it would eliminate one point to failure
One could suggest using a MOARstruder to get better layer bonding (but maybe losing on teeth accuracy)
Or maybe it's just that I want to see MOARstruded NylonX :p
3D Printing Nerd
After watching this I am impressed but I did notice somethings that went wrong and if you want I have my advice.
Where you went wrong is three part.
1. You separated your print during the process and added in the nuts.
This could be solved by placing the nuts in after it reaches 1mm rim on the hole without stopping the machine. I have done this with mixed results. I'd you do this Put adhesive (personal fan of Gorilla Glue) on face of nut and on top of nut.
2. You stopped the machine and didn't reheat the material. I don't recommend reheating but it is a possibility to a solution.
3. You could mount the go pro under the back for safety.
Last thing you could do is a two part piece that has separate bolts to attach. Because the material held up just not the seem. (Counter sink The nuts for this method as not to increase the size.)
or redesign it to be a 2 piece from the start with interlocking teeth instead of being flat on the two surfaces you want to stick together. If you also have the nuts half way into each piece it would make it a lot stronger
The 3d printing community needs more of you guys, testing the limits of the materials and the process. While the result seems obvious It revealed data that will be useful for anyone pushing the envelope. Great job.
"You wouldn't download a car"
*Its getting closer*
Printed parts are getting harder.
And so am I.
*Cue music*
You wouldn't print a lady's purse...
I would still love to see a retest with a one piece gear.
He also needs to print It standing or at a 45 degree angle to keep it from "delamination".
Stuff this simple should never be printed.
And use brass inserts instead.
@@lucaslittmarck6804 problem with printing it at a 45 deg angle is that some teeth of the gear will be prone to de-lamination.
no good way to print it on FDM printers.
@@darkshadowsx5949 Non-planar printing with variable layer height might help.
This joke is so good, 3 years later and people still don't get it
Please come back to this, it's extremely awesome seeing products that don't belong in a car get used
Not watching to see if it fails. Watching to see how spectacularly it fails.
BikerTrashWolf wasting plastic. Better do a mold and aluminium casting.
Please don't use WD-40 as a grease. WD-40 is a mixture of grease and degreaser, it is not good for greasing up parts
BrainSeepsOut, yup, just some 30W oil should be fine
Yup, just some 30w oil should be fine.
For gearboxes and diferentials mostly are used oil with SAE above 75W
it doesn't have any 'grease'... nor is it a degreaser.. .its a WATER DISPLACEMENT solution... nothing about it is a de-greaser, rust remover, anti-seizing agent, or anything. Its literal purpose is to displace water around whatever you spray. It just so happens that the displacement mixture also penetrates, and thus helps with rust, but in-so also CAUSES rust... its an oxidizer basically... it also is a HUGE enemy of ANYTHING 3D printed, because of its penetrating nature...
As a side note, something like Brake Cleaner is in fact a de-greaser, sooooo
kwestionariusz1 75w- gear oil viscosity is comparable to engine oil of a grade somewhere between 10 and 20. The reason gear oils have a higher grade is so you don't mix them up with engine oils.
Try again but try this: pause the print, position the nuts and use a soldering iron to quickly sink them, use extra long bolts that extend appropriately in gear side as far as possible. You could possibly reuse what you have printed by drilling. You will need to create holes for the bolts either way. Or you could possibly add an array of smaller bolts to do the same between the existing bolt holes. The main idea is to use bolts to resist shearing and separation. JB Weld probably could be used on mating surfaces.
Do it again with press-in inserts.
Also temper it if possible.
Might need to 'cure' it with some slow heat and cool cycles, to fuse all the layers.
I'd say do it again man, this was interesting. :)
The ring gear you are trying to replace uses a case hardening steel with tensile strength probably between 500 and 1000 N/mm^2, no way in hell a 3D printed, non-homogeneous plastic part with questionable layer to layer bond can do the same job.
Stefan. Exactly what I was thinking. I guess we are both mechanical engineers :)
Стефан Парийски I mean no fucking shit that's not the point of the video
Even if it was injection or compression molded it would get destroyed as soon as it was on the pavement....maybe a super tough nylon material with a high steel fiber content but unlikely
Success is not the point of the video. The point is to see how far away we are from being able to do this. IE is it say 3 orders of magnitude or 2 or even 1? Ordinary nylon I would guess would be 3 orders of magnitude too soft. With the carbon fibre depending on type I bet it's possible to get within a factor of 25.
Under ideal circumstances it might last a few minutes of slight acceleration. One turn, or flooring it, or reversing it and it's toast.
With how the thing snapped, it just seems like the design could use revision instead of throwing the idea out.
If the teeth were all chewed up instead, I'd be concerned. Hey if MH is willing to supply you with more filament to do it, why the hell not?
I guess there must be a reason why a truck differential ring gear is machined out of a solid piece of steel. But I love that you guys tried this! It's how you learn!
Simple fix: Print it the other way up. That way the torque transfer from the gear to the bolts is all one continuous print with the layer holding the bolts in being under compression during use due to said bolts.
Not sure I understand this, could you explain please?
I'm actually amazed it even held up that well! Kudos!!! I can't wait to see the revised version.
Yes, please do it again. I'd suggest Threadserts. Larger OD to spread the load out, threaded to allow you to loctite them in. Adhesion to the nylon shouldn't be an issue as the loctite is just to wedge the Threadserts in place.
Like this 7/16-20 LH thread helical insert.
www.threadtoolsupply.com/716-20-lh-helical-thread-inserts.html
John Pickens: I think helicoils (for stripped hole repair) could do as well a job if he were to make it as one piece. Once you tap them in, they don't come out, and provide a strong thread face to hold on to while supporting the shear-load.
Helicoil, Threadserts, same thing. Look at the link I posted, same stuff.
I saw the title and thumbnail and literally LOL'd at the idea. But I respect the ambition and doing it for science.
I would like to see a part 2
That is some sick skills, putting back the previously printed model and printing over it! Awesome job!
DO IT AGAIN!!! This is really cool!
Im just a hacker who loves working with computers , cars , drones and industrial machine's.
I always wonder about stuff like this, its so cool that your making videos like this and allowing people like me to learn .
Thanks 😊
Try again with a different design!
I would love to that thing working properly. The material seems to be strong enough to handle lower torque and lower speeds. By the way, could you try to make a transparent cover to protect the differential? :D
I would retest with threaded inserts. That's what's used in the composites industry for things like medical tables and aerospace components. Just drill and set with an epoxy. Also, a side note for JB Weld, heating it to about 120 when drying helps with flow and adhesion (good for getting out voids), but will make it runny until it starts to cure. Also, for a lubricant, you could try graphite.
You should definitely try it again!
as soon as you showed your method for getting the bolts in I had a hunch that it would fail there, still interesting to see and defintely the right material if you a 3d printing something like this.
You should perform annealing to the finished nylon CF part in an oven set to an optimum heat settings for nylon to strenghten it....
I am well familiar with how differentials work, because I have to build them more than a million times in Lego.
Dude I love legos, even got a 5 speed trans built out of technic parts with reverse and standard H shift pattern.
I'm going to start casting alum lego gear parts, you reckon there was a market for em?
Patrick Baptist Maybe ☺
When you are done printing, bake it. It will strengthen the layer adhesion. Also, try MarkForge.
the printer isn't the problem. ungodly amounts of money on a printer isn't going to improve quality to godly proportions..
people seem to have a problem understanding that the more you spend doesn't necessarily mean the product quality improves.
you can spend a billion dollars on a FDM printer and still have issues. you will over pay just make the same part because of your status ego.
you could print the gear and use helicoils, or a slot on the side and slide the nuts in
Came here to suggest this as well.
ok I'm a 3d print idiot but if this materiel is as strong as he says it is why can't he just print the threads into the part? I mean if it can handle the torque of the motor why wouldn't it to be able to handle the torque from tightening the bolts into the "plastic"?
You have to try this again. For science! Seriously, this could change the game for a lot of people trying to mesh their love for cars and imaginations with a 3D printer
It could not. Gears are usually made to last for a very long time and to do so, the stiffness and bending strength of them are pretty critical. The bending strength at the tooth root are very much lessened because of the 3D printing and even with chopped CF, the stiffness is much less than steel. It will not be a usable part for a car.
Sebastian Jørgensen Do you even watch this channel, bro? These guys have shown that even PLA does amazing things as printed gears.
Sebastian Jørgensen: I don't think the idea was for practicality, it was just to see how long it would last. Failure was assumed from the beginning, hence the beater pickup. You are quite right in every other way though.
Claws61821: I do, but I am also an engineer who has had the task of designing a gearbox for a racecar. Those gears are heat treated to have strengths much well above 1000 MPa wheras 3D printed PLA had a strength of 30 MPa last time I put a piece of it in our tensile testing machine. For the wear on the tooth flank, the tooth flank stiffness is also very important as the steel tooth will indent into the plastic tooth, thus not keeping the involute shape. This exites fatigue failure in the tooth flank but also gives higher losses and more wear on the plastic tooth.
I know that this is not a racecar and the gear that it replaces is most likely a hell lot cheaper than those we use but even if it is simply C40E (Edit: type of steel, commonly used for gears) there are still a factor of 10-20 in difference of strength.
Especially because 3D printing does not give very homogenious parts, the fatigue strength is probably also a lot lower than for a machined plastic gear so it will never be suitable for use in the transmission of such a heavy vehicle. A 12V childrens car is probably a suitable application for 3D printed gears.
Philospher of Nonsense: Zac Parks sounded like he imagined that it could be an acutal usable part for this car if it didn't fail where it was paused. 3D printing is great, but it is not a wonder where you can replace anything with plastic and especially not when the manufacturing process of 3D printing results in so much weaker parts than other plastic parts.
Go for it man. I want to see if it able to withstand high-stress testing and normal driving conditions.
please do it again!
Thanks for the experiment and the cool video. Since the main problem is not the material strength and rather lack of enough layer adhesion, I would suggest two remedies:
- Printing at an angle. probably 30 degrees or somewhere there.
- Putting the printed gear in a clay/plaster/silicon mold then heat it up in the oven for a couple of hours so the plastic remelts and the integrity improves. And the mold can be removed using a water bath or just physical force in the case silicon.
I'm really curious to see the results and I hope you consider them.
Cheers
Wait. Couldn't you just put slices from the center to the place where the bolts are supposed to be? Then just slide the bolts in, see how it goes? Forget the tap and holes. Just make sure you use the nuts slid into it. Benefits are that it's all once piece and it should retain most, if not all of it's strength.
Please try it again.
^^ to the top! ^^
Nathan: I was just thinking the same thing as well, where you would slide the nuts into galleries arranged radially, however there is an issue with that. Due to a relatively imprecise fit, it would put "pressure hotspots" on small areas of the ring as a structure. You would want to spread the load as evenly as possible. Quite literally how the tip of a nail can pierce more easily than the head, if you get my meaning.
Thanks for posting this man- I’m actually in process of creating small precise gears requiring torque, but not at this level. This gives me a boost of confidence!
More!! More Gears!!!!!
You have just earned yourself another subscriber love the production of the video
I wanna see if it works
i couldn't figure out if this had been revisited or not, but here are some tips for revisit:
1. don't print it flat, tilt it slightly so that the shear force isn't directly applied to a single layer
2. find out if there is a curing process available, might be able to heat it to a certain temp for a certain time to cure it for more strength
3. don't pause it of course, just model the threads right in - should have no problem mating with threads that large
4. get/make a clear cap to put on the diff and use proper gear oil, don't just let it hang out in the air unlubricated
5. someone else mentioned using a go pro, great idea
WD40 is a terrible lubricant. Just fyi. Basically anything is better than wd40.
also dissolves adhesives
backtrackit anything with any oil content at all dissolves adhesives. Next time you pull a sticker off something and it leaves sticky residue, rub a finger on your forehead and rub the sticky spot. Even the oil from your face will dissolve adhesive. Not really a special wd40 trait.
backtrackit: WD-40 is nothing compared to PBlaster, especially in penetration. That'll melt a styorfoam cup very rapidly compared to WD-40, which is just a waste product of gasoline manufacture.
As a lubricant its only ok. It is a penatrating fluid, degreaser, rust remover and a few other things
AvE compared a bunch of different penetrants, and WD-40 turned out to be one of the best ones. And PBlaster (PB'laster?) actually did pretty badly. This was with a sample size of 4 or 5, so take that for what it's worth, but I was surprised to see WD-40 do so well.
Please do it again this is such an awesome idea it deserves at least a second try.
Yes, do it again!
The fact that it moves at all is impressive. There are enormous loads in a rear end.
I embedded nuts into a handrail once. I haven't skated since then.
Belias Phyre Don't have the balls to do it again eh?
After medical bills, I didn't have an arm and a leg.
You are living in the wrong country if you have that kind of medical bills.
Got a new kidney inserted, and it cost med about $30 for nearly two week hospitalization. And yes, it cost me about £250 a year for medication and regular medical check up. No, I don't have any personal insurance for this.
You're right, I live in the wrong country. I'm a refugee from Kekistan. This isn't my home, though they've been gracious to host me. As for the cost, I'd rather lose an arm and a leg rather than my soul to a commie totalitarian government.
No need to lose soul, leg and arm to get descent health care for about the half price of USA.
Maybe print it as two separate parts? Kinda like a lock & key type design so you can put the nuts as well as lock the parts together? I apologize if I'm not explaining things properly but hope it gives you some awesome ideas!
1:57 thumbs up for cat
3:52 There's a glitch in the Matrix...
Yeah I don't think it will work with a redesign. But you are in it now so keep on keepin' on!!! Oh BTW I love the shop manager that came out to inspect at just the right moment (talkin' about the cat!!!)
Do it again!!!
That’s a ton of torque. I can’t imagine the strength needed if you really get into the gas petal. Good luck on version 2.0!
WD-40 is a solvent!! Not a lubricant. Use a lubricant that won’t damage your print!
And not use that much oil. Some on moving metall parts and not half a liter!
WD40 is 1 part motor oil, 1 part fuel oil. It is not a solvent.
definitely worth another shot. don't pause the print! Very curious to see what happens
Sprayed WD-40 in there... nice... 🤦🏻♂️
my 2 cents is that the nuts were moving around from the shear force on the bolts. the nuts moving around helped the part split where the two halves were joined
Coolest thing ever!! Theoretical turned to practical. I actually got goose bumps!!
I think that custom threaded holes would ensure the layers properly adhere. But, if you are going to stop half way and put nuts in, maybe printing it in two pieces and securing them together with extra bolts?
Do the spider gears as they wont need nuts or threading
21babydew: The splines would strip quite easily though, or do you mean only the spiders on the carrier?
Just the carrier spider or shoot try it all like thats what this channel is about
Put the custom threaded holes in place. I believe that is exactly where your issue resided. The material seem to hold up quite well aside from that.
I've used NylonX for years. It's a great filament for anything structural or anything with purpose. Pla or similar material are just for concept work.
3D printed parts are nowadays still anisotropic saddly... not really enabling you to print hyper strong parts whatever dirrection the stress is applied to...
Still very inspiring ! I guess you must continue doing this kind of videos !
unless you use the extremely expensive M1 carbon printer...That thing is one cool overpriced machine...
You have to admit though, even after 2 failures, those gear teeth really held up well.
I'm completely amazed how well the teeth held up even after the failures. Multiple. I was expecting the pinion to strip the teeth on the ring gear immediately. That makes me think that you should skip the hex nuts, and carefully thread the holes directly in to the ring gear. Use a course thread tap, and bolts. Also, go ahead and assemble the back cover, and fill it with gear lube. Go easy on the clutch! I wouldn't recommend doing "burn outs" with the plastic ring gear. If it doesn't de-laminate, it could surprise us all.
MOAR!!!
Yes, totally to a full print and tap out left handed threads directly into the final product.
Also print it with slightly under sized holes so tapping out the threads will be easier.
Either design this as a continuous print, or as two different prints with teeth to increase surface area.
Also WD40 is a terrible lubricant
Suggestion: step 1 - before printing, set the nuts prethreaded onto the bolt shafts at the desired hight.
Step 2 - set the bolts on the head with the prethreaded nuts shaft up in the pattern desired.
Step 3 - set printer to start printing at the correct measured hight on the bolt shafts in order to embed the nuts into the structure without having to pause the printer.
In this method, it is all one piece and you can still remove the bolts because you have access to the heads.
Maybe just reheat the part before restarting the print
Michel PASTOR I was thinking this too, use a non contact thermometer to read surface temp at the pause, then heat gun or hair dryer to bring the surface back up to the previous reading.
Although I think using slots to slide the nuts is may work out better.
You should try to print it as a single piece with slightly oversized holes, then use metal inserts for the threads. You basically push them in using a soldering iron. You get 90% of the benefits of threads, without having to pause your print, and it's much better than just cutting threads into the nylon itself.
Wd40 is petroleum-based. It melts most plastic.
Nylon is extremely resistant against most chemicals
I've never seen WD 40 melt plastic. Any type, ever.
Maybe that is not the prpblem, but never use WD40 as a lubricant!
We use steel helicoils in a dense plastic called Delrin at my work. It holds up really well.
WD = water dispenser
WD lubricant
Usually, when you change the material you use in any kind of part, you must redesign it in order to adjust it'resistence to the new material. If the design keeps the same, probably, it will not work in most cases, but with changes, I believe you will have success.
the WD-40 will destroy glue and plastic
you'd need to autoclave it, the layers are strong against their tensile, but applying torque that isn't perpendicular to the layers, will cause separation since they aren't actually fused.
The fact that you was able to back out of the garage is amazing!
Slots to the inner edge of the ring, slip the nuts in via the slots and bolt up, that way the whole thing can be done in one print, you might also want to print it at a 45 degree angle so the grain isn't horizontal to the rotational pressure
Pretty slick! If you need to embed the nuts rather than drive them in from the outside I would suggest designing keys and slots into the two segments of the build. That would channel the torque load better than the flat surface of this design.
I have one thing to add, i have used 3d printed parts in production and for my part it's not the filliment that is the biggest issue. it's delanimation and layer density in extruded printing. But if you need it to last say a 100 000 handelings before it brakes extruded is out of the picture. The only thing that works is SLS. SLS PA have been used in production for two years now and it is only chipped in some corners. and this is where aluminium is normally used but it doesn't work due to weight.
I'd say try encasing steel pins through the width to prevent seperation, but really cool idea. I think it could work, just need to keep at it!
I'd love to see you guys try this again in PEEK!
Just an idea, you should put structural pins in both sides of the ring gear. That will help with the lateral force on the separation
I think you have to right idea with the nut inserts. I doubt that material could hold up to the stress of being torqued down.
Maybe try stopping half the depth of the nuts to insert them and keep your polymer warm somehow. That should help spread the shear load.
An exotic additive lathe 3D printer would be really good for this. Then there are no axial layers but only radial ones. And there are much less shearing forces because the meshing teeth distribute the load well.
Print it at a 30° with supports, so that your z-axis does sit perpendicular with the torque being applied. I’d also use press in (with heat) corrugated nut inserts so that you’re not pausing the print process halfway through.
As a drivetrain rebuilder of 30 years I was interested to note that there did not appear to be any check of the contact pattern between the pinion and ring gear..they are lapped in as a matching pair during manufacturing and cannot really be mix and matched successfully. Poor contact will result in noise at best and tooth failure in short order. I appreciate the effort made in what they did..and I won’t dismiss it as potentially feasible however the stress on gear teeth requires strength that I can’t see being achieved with a printed part.. but I’m not involved in that industry so maybe it is possible. I enjoyed the video and thanks for making it guys.
I'd love to see it running. This time try to create treads inside nylon. Not necessary dense but large treads that are printed. Alternatively leave holes and tread after printout. I think that it may work...
Proof of concept! next project crank, rods and pistons.
YES! Round two please! This is a bonafide rad project, thanks for the video guys!!
I absolutely think you should redo this video without layer issues. The shear gear in most transmissions is plastic and it holds up just fine until abused.
I definitely think you should revisit this again. Taping holes isn't going to be an Issue. Making slots to slide in, I think would be less material and less strength , though easier than taping; but what I'd like to see is all the gears in the differential replaced with the carbon fiber filament. If this gear can hold up to the strength pulling the truck, I'd love to see a closed and filled differential running in second gear on the road. At that speed both gears should heat up, and of they are the same material, expand at the same rate. I'm curious how many expansion cycles it can endure at low torque before they self destruct or have the mating surfaces degrade before the teeth along the ring sheer past each other.
This satisfied my curiosity, although it failed it was needed to be tried
Yep drill and tap holes maybe even get a junk carrier where you can grow more holes for more bolts to help spread the load out between the carrier and the ring gear
I love the JB Weld optimism
Trial and error. Awesome demo. I would make a base for the printer that would hold each nut up to that 17mm so that the print can go without stop
Ok, open up the nut hole tolerances so you can slide them in with adhesive while still on the print bed with a quick print pause.
Anneal the print afterwards to maximise layer adhesion.
The direction of force is always going to work against this inherent weakness in printed parts, particularly in sheer force, though doing the above will give it the best chance.
1000!!! Do this project AGAIN!!! THIS IS AN AWESOME IDEA!!!
Yep, use fusion 360, it has built in thread generator, you can match those left handed bolts as designed.
I was more concerned about the fine thread getting any grip on the plastic that anything.
Gear Down For What? Worth a try with a test print.
i doubt the resolution would be hight enough to get it to work without chaising it with a tap first.
kleetus92 I don't know how fine they are, but when you can print 50 micron layers, I think any threads are possible.
Oh.... I didn't think it would go down that fine. Even so, chase it with a tap, no biggie.
Can you print it solid and then drill and tap it? Will that material cut that way? I love the idea and can't even imagine what a game changer it would be to be able to 3d print auto parts.
Have you tried printing in 2 separate pieces? Make the top part with dowels and the bottom with corresponding holes. You can make some double sided dove tails to connect the 2 halves and laminate the whole thing with epoxy. You'll have to compensate your scale to accommodate the resin but should solve your problem.
Probably drill out little pits before continuing printing so that there is some pressure points and you could put some adhesive to the pits so the two parts sticks well