I'm a CNC designer, and I can tell you that make the nut rotates is the right choise on a long and/or fast screw. But you will need to constant lubricate the nut manualy. Normaly when we use a rotate nut, we make a much more complicated assembly so we can make a automatic lubrication. Very nice work. Cheers from Brazil.
@zomgthisisawesomelol It's a good option on long and fast travels, but you'll need a reductor (with low backlash) on the servo to meintain the milling force. For large portal machine you ussualy need 2 servos, one on each side. And the mounting system needs to keep a pressure from the pinnon on the rack, the eliminates the backlash from de rack. Cheers
Lubrication isn't that hard just take the seals out of the ball nuts inside the bearing housing and have an auto oiler drip oil onto the screw. As for the screw being a bit bent they are pretty easy to straighten, all you need is a set of v blocks a dial gage and a long pry bar or hydraulic press, you just rotate the screw on the blocks find the high spot and bend it back with the press/bar rotate again and measure and keep repeating until you get it straight. Just make sure you use a block of wood or rubber between the bending implement and the screw so you don't mess up the screw surface. As for the pitch you ideally want 10-15mm on a gantry that size. Looking at those angular contact bearings being grease sealed they probably top out at 6-8k rpm I'd guess. You probably don't want to run the ball nut much more than 3k rpm it really depends on the ball size and recirculation path though, longevity will be hugely dependent on adequate lubrication and keeping dirt out, there is a company out in California Called balltec that can supply balls in any size you like down to the micron if you need them, you could potentially polish the internals of the nut then reball to get them running more smoothly, I have rebuilt a lot of ball screws, many people treat it like a secret art but to be honest as long as you don't need to have them reground you can do everything yourself. The main reason for a stationary nut is so you can flood oil lubricate it to push contamination out! That alone leads to improved longevity even if the seals are non optimal, typically most vmc's pump a shot of oil into the screws and linear ways around every 8mins then they have an oil skimmer in the coolant tank to remove that non soluble tramp oil from the coolant.
Please make videos about ever aspect of cnc design. Professionals are not on UA-cam; ending with lot of miss information on youtube. Can you recommend a book about cnc design.
Very interesting way to turn the problem on its head. Makes me think of a helical rack driven with a gearbox. Except unlike the geared helical rack you’re not having to deal with (much) backlash.
Dont get why even spend money at ballscrews. Just use a normal 20mm threaded rod with 1mm pitch. Screw on it 3 or 4 or 5 screw nut and tight them together with hand. Then weld all together ,put in lathe and make it nice .The result is 0.000 flashback and always accurate. Will bend a minimal if so long threaded rod, but is doesnt matter while you turning the nuts and using the linear rail.
Was thinkning of upgrading the work 5x10 router to this so it can handle steel, now makes me think that I'm gonna have to put in a lot of care into designing the rotating nut mechanism and finding good screws... great vid!
The rotational inertia should not be limited to the mass difference of the nut vs screw, but how much of that mass is distributed away from the center of rotation- I suspect that the rotational inertia is not as much of a design constraint as you may think- the inertia of the gantry will obviously be more of a factor, so being able to put the motors on the table rather than gantry would be nice. And as another person posted, spinning the ball nut will complicate the oiling system. But there is never a right way or wrong way- very nice work and good job thinking outside the box!
Nice. I've never thought of solving the whipping problem that way. I built a 3m x 1.6m cnc router 12 years ago and it's got 3m long lead screws with recirculating nuts like the ones you had on yours. I do get whip quite badly when the gantry is at either end and it's moving at full tilt. Full tilt is about 100mm per second. However the pitch for mine is 10mm so it's not as bad as yours would have been. I've learned to live with it. For the person that suggested using a chain or a belt, the reason that I went with lead screws is that it's so much more accurate. I can make watch parts ....or I can cut out kitchen units. Anyway, thanks for sharing.
You can use way smaller ballscrews since your spinning the nut. I designed a similar machine but ran the ball screw though the center of the steppers I'm using.
I've never seen that idea before, this is awesome! Also if you didn't show you welding, i would have never guessed that you built the pulley shaft from scratch
My 0.5" X axis acme threaded rod is over 6 feet long and I had never thought of rotating the lead nut rather than the rod (rod is hooked to stepper currently). I think between whip and rotating mass reduction, you have a great idea! But reading comments below, using a stationary chain and driving sprocket is a great idea too.
Hey Christian, thanks for this. The chain and sprocket would work well for low precision high-speed applications. Not so good if you want to mill aluminium to a decent accuracy... Too much backlash and/or inaccuracy in the chain.
@@FloweringElbow4SuperNerds Good input. I will go for the rotating lead nut; simple to do in my case. simplecncrouter.blogspot.com/2010_04_11_archive.html
Inductive proximity sensors are very repeatable, dirt, wear and corrosion resistant and not particularly expensive. Dust will likely cause a pogo pin to jam over time.
The very best combination is inductive sensor then encoder zero index pulse. I've done inductive only and zero index on 2 different machines, zero index is dead accurate and I know I can resume a job the following day with the exact same homing position.
@@SuperYellowsubmarin Hybrid steppers will give you that but unless you're getting missed steps then once homed then any machine is going to be that repeatable, I only have to re-home my machine after a power down or crash. Inductive sensors are repeatable to a level far higher than most people need
@@mfx1 I mean encoder index as a means of homing ! Centroid Acorn allows that. Of course for a router it is not needed, bot for a mill or lathe I actually noticed the difference.
It's nice to see the inner workings. Langmuir systems is coming out with a new 4x8 plasma table that uses this method for driving the gantry. First time I saw it I thought it would be a genius way to handle screw whip. Anyways good luck with the build, it looks awesome so far.
First thing I thought about when I saw your layout is the issue you are going to have with chips getting into the threads of the shaft. Need to look at placement of the screw on the underside or at least on the outer edges where you can put some shielding over them. Also, definitely new two shafts with that placement or one underneath centered. Keep up the good work
Hi Chris. Thanks for sharing your thoughts. There are two ballscrews, one each side... They are situated on the outer edges, but you are right, we'll need to setup some kind of shielding.
Hey Joel, thanks 😊 Recon we could fill the housing with grease. Still working on how to seal the dust and chips out of the screws though, so until then, I think it might attract contaminants like a magnet!
Very nice work! It looks like the problems are not coming from the rotating nut arrangement (which to me is genius, good choice on your part!) but from the too-fine pitch...
Hi, I really enjoyed your video series in making your industrial size CNC. I am wondering if you have the video of the process in making the rotating ball nut. I watch what you had posted carefully and wanted to know how you got the gear installed. Thanks... BZ
If you are using actual servos many have a hard stop homing/stop by count so you might want to check more advanced settings...I hard stop and home by count.
Thank you very much for this, i really like this idea - gonna look at something like it for my plasma table build instead of fighting rack and pinions. :) (but using 10-15mm p itched screws) :P
Great design. I suppose the only disadvantages I can think of are that the motor has to sit on the X gantry, which increases it's inertia (although probably not significantly at all) and that the additional wiring for these motors that needs to come off of the moving parts, meaning more complexity and more vulnerable.
On the other side you do not have an inertia problem with rotating the ever longer screw. THAT likely is a lot more intertie than the ADDED weight of the motor starting with a specific length.
@@ThomasTomiczek I think it would depend on the specific configuration. I think its also the kind of problem that you dont know the ideal solution to unless you build it and test it. Regardless, I doubt that the inertia is the limiting factor, especially because CNC machines generally run pretty slow. Mechanical bandwidth is probably more than high enough.
@@tHaH4x0r There is more than just metalworking cnc. A woodworking router cannot be called a slow cnc. The "holzher ecomaster 7113" cnc also has the design of a ball screw. Nevertheless, the speed is 60 m / min at high speed (g00), 3500 mm in the X-axis direction. True, later models were already made with a "rack'n'pinion" design because they are cheaper. . .. At least I think so.
call me crazy, but what about a combination of both? nut rotating twice the speed of screw makes for a 10mm pitch. much faster rapids. conversely, driving it at half speed gives 2.5 or something. higher resolution. may need some fancy software to morph between high resolution to coarse as it feeds and rapids... then again, its just differentials... easy to mentally picture, hard to write as an equation... :D
Thanks for making videos on the rotating ball nut design. There’s not much info about it. Secondly if the motor is rotating too fast change the pulley confirguration to get a higher gear ratio. Less rpm for the motor and more rpm for the ball nut
Hey Jamie, thanks for watching ;) The real problem I have is too much rpm at the ball nut. In order to get good speed the ballnut's got to turn too fast, and it wont last... I have some new ballscrews with a larger pitch now ;)
I have a question about this. Since there are limitations with the top speed of the shaft rotating when very long, and there are also limitations with how fast you can spin the ball nut. What if motion happened by rotation the shaft at some arbitrary rpm and then you also rotate the ball nut? So speed would then be multiplicative.
Yes it is possible, but very complicated. The relative speed is the important limitation (how fast the ball bearings in the nut are moving round their helical path against the screw...
Hi friend... I don't know if you read these comments... I actually enjoyed your build.... although, I am a designer, and I want to put this into CAD... specifically SolidWorks V2022. And I have a few changes I want to make. I would enjoy to benefit from some of your experience... such as what size your ball screw is, and where you got your parts. What size Servo Motors, and where did you get your controllers and other electronics. I am a mechanical designer, not an electronics guy. That is why I am asking for such information. I hope you can help me. Have you put together a materials list and where to get the materials? Also some tips about setting up and software you are using for your setup. I do have access to CNC Mill, and a nice lathe and surface grinder and other equipment so I can do this project. So... if you would be inclined to assist me I would really appreciate it. If your willing to assist me... I am willing to provide you with my designs, and also the plans I make. P.S. setting up the electronics is going to be an issue for me. I can put together components to build computers... so I do have some experience with playing with electronics. But I have never worked with controllers and servos. I can be reached at... I am Leslieallen inventorshaven@proton.me Please contact me. Friend... if you have Signal Messenger it would be a big help, that way we can easily communicate using video or messaging. I prefer video actually. Anyway, write to us if your willing. Thank you. I will post this on each of your pages... so hopefully you will see it.
Trying to think outside the box, could you put a motor on the ball screw as well as the ball nut? Leave what you have for " fine control", but also drive the ball screw in the opposite direction for fast movement? This might be overly complicated, and not worth the effort, but might save you the money of new ball screws.
Hey Shale. That's a really interesting idea. The problem I have is that this would not reduce the speed the balls in the ball nut travel (instead they would increase proportionally with the added speed of the ball screw.
@@FloweringElbow4SuperNerds Yes it would. I did not fully know if your issue was the speed of the ball bearings themselves, or you might be having an issue with how fast the ball nut its self was spinning causing the issue. By splitting the driving force you wouldn't have to drive the ball nut it's self as fast. Yes the speed if the ball bearings would be the same, but may reduce the issue. Again, a pretty overly complicated solution, but I don't know how expensive those ball screws were, and may be an idea if they were expensive. Of course the idea has its flaws and may not be a good one
Hey, I'll let you know when I get the new ones installed and run it a bit ;) But yeah in general, I think they can be a touch better for those things, but all depends on the details...
Great stuff!😁 I'm considering a rotating ball nut design on a 1.5m long axis to avoid screw whip. One thing I can't figure out is how to lubricate the nut; as it's spinning I can't just add a lubrication line! How are you doing this on your build?
A the moment I use the oil nipple port on the ballnut when its stationary. That or I just put some oil on the ball screw itself. Crude, yes, but it works.
@@FloweringElbow4SuperNerds Thank you! That's pretty much what I'm doing with my current router; built before I knew about building in lubrication systems...!😁
When you were calculating for the inertia you didn't account for the mass of the carriage itself since it is attached to the nut assembly, not sure if this is thr best choice. Will be interesting to see how you solve this.
If I understand correctly what you are saying, the mass of the carriage remains the same whether you are rotating the screw, or the nut. The inertia involved here is the comparison between spinning the nut, or spinning the screw... the balance of the inertia remains constant either way.
hola. Cometí el mismo error en mi CNC, si ves videos de maquinas profesionales (Biesse) observaras husillos de paso 50mm/rev. Si piensas actualizar a motores de bucle cerrado te daras cuenta que en altas rpm pierden hasta el 80% del torque. Hay que hacer que los motores trabajen a baja rev donde su torque es mayor. Tambien ayuda usar tpreload double ball nut.
Hola, gracias por tu comentario. Es bueno saber que no soy el único ;) Estoy usando servos AC de segunda mano, que parecen estar funcionando bien hasta ahora. Obtuve tornillos de paso de 20 mm al final :) You can see it in action here: ua-cam.com/video/obR-l7qVhQY/v-deo.html
Excellent setup. I read that you increased your ballscrew pitch to 20 mm. Was this because you were on the limit of your servo motor RPMs? I'm on the final stage of a CNC router too and I'm considering purchasing 10 mm pitch ballscrews and using rotating nut too. Ideally I should get 20000 mm/min easily with the servos but I would love to know your experience rotating the nut that fast. Thank you :).
Hey Leonardo, thanks for getting in touch. Yes I now have 20mm pitch TBI screws (see them in the latest vid). The reason was that I the ballnut has a realistic max speed of about 1000rpm on these cheap screws... More than that and they make a lot of noise - and wont last long! The 20mm ones I got have a different kind of recirculation path for the balls and are designed for higher speeds - but still have a max of about 2500rpm (but travel should be plenty fast enough a reasonable amount below that speed with 20mm pitch). Are you planning rotating ball nuts or to spin the screw?
@@FloweringElbow4SuperNerds I'm planning to use rotating ball nuts as you. I'm double checking now if the ones I'm going to use have a RPM limit close to 3000 rpm which could be a problem because I really need a 3000 mm/min speed to be comfortable.
@@leonardomarsaglia It depends a lot on the diameter of the screw I think. Mine are a bit large at 32mm - even ground expensive THK ones have a max speed less than 3000... Best check the documentation. What make is it?
@@FloweringElbow4SuperNerds Sorry for the late reply. The brand is PMI. But I could find some 25 mm diameter 25 mm pitch screws and now I'm finishing the mounting setup for those. Thank you :)
Design Option 1: Steel Rod Rotating at 100 revolutions per second Moment of Inertia of the rod (I): 0.40625 kg·m^2 Angular Velocity (ω): 628.32 rad/s Rotational Kinetic Energy (E): (1/2) * I * ω^2 Result: The rotational kinetic energy is approximately 80,562.5 joules (J). Design Option 2: 10 kg Object Moving at 70 cm/s Mass of the object (m): 10 kg Velocity of the object (v): 0.7 m/s Kinetic Energy (E): (1/2) * m * v^2 Result: The kinetic energy is approximately 0.245 joules (J). Conclusion: Comparing these two design options reveals a significant difference in terms of kinetic energy. The steel rod rotating at such a high speed (100 revolutions per second) possesses substantial rotational kinetic energy, approximately 80,562.5 joules, due to its moment of inertia and angular velocity. On the other hand, the 10 kg object moving at 70 cm/s has much lower kinetic energy, around 0.245 joules, due to its lower linear velocity. This difference in kinetic energy is crucial when considering the design and application of these systems.
Kind of show this in this video: ua-cam.com/video/k47LBPEemvU/v-deo.html Essentially with four set screws per side for alignment, and threaded at the end to tension the screw.
ball screw needs to be in the middle. or there needs to be 2 of them, one on each side. ive got a professional built laser table and a home built plasma. been through this a time or two.
this is fun but insane, rather use a rack and pinion, or just a stationary chain snaking through a u-bend over a driven sprocket in place of your rotating nut.
I've seen rotating nut setups before but never used one, they always looked the business. I once read that you should tension the ball screws too on a setup like that. What's puzzling me is what pulley are you using on the pipe? I thought it looked like a taper locked pulley but after closer inspection I noticed it isn't. I can't figure out how the (driven) pulley is fastened to the pipe.
Hey Paul, good question. The pulley there is actually bonged on with locktite. It works to hold it, but a right pain to remove the pulley if you want to change it - I know!
@@FloweringElbow4SuperNerds Thanks for replying... That Loctite is seriously strong, to remove it we used to get a blow torch on it. I might be wrong though. Did you buy a regular pulley and turn the middle out on a lathe or did it come with that bore? I'm curious to know as I've been after making a rotating nut setup for a while.
Hi can anyone tell me if is possible to have a play in the nut if I rotate the screw manually not one full turn just a little bit let's say 20 percent and the nut won't move is that normal ?and what could be the cause tnx
hi i have a cnc router i need to change the motors of a 3 meter x 2 meter router, and i was considering buying absolute encoder servo motors with standalone gsk controller, can someone guide me, is this feasible? or only servomotors with incremental encoder would be enough.
Hello Hanz, I would go with 17 bit absolute encoders all the way. You can tune your servos perfectly with that. If you are planning to purchase chinese servos, you'll find there's not much difference in price between 2500 PPR and 17 bit absolute. I'm using delta 17 bit absolute servos on a lathe and they are wokring flawlessly.
Having a rotating ballnut might not be much better than a static one. The inertia is rotational inertia which means that it depends on the mass distance from the radius, for your monster balscrew it will help but for smaller screws the difference will be less. Also the whole machine gets more inertia as the motors travel with the gantry. Lastly you also have to have bigger cable chains for all the extra wiring. That being said, for chinese ballscrews which might not be straight this system minimizes vibrations and the risk of the ballscrew throwing around.
My brain isn't a fan of that. Most of the rotational inertia would be very near the centerline, so I don't know how much of those 15kg actually matter - and I'd wager the inertia of the 2 other axis probably dwarfs it. Also that just seems like a great way to centrifuge oil away from the ball bearings (especially where they meet the screw, which is most critical). Mostly though, that just sounded terrible. I'm also not sure that belt bodes well for precision since the encoder is in the motor - I'd wager you're losing a few steps of precision when switching directions. If it were me I'd probably just rotate the screws, at least until I got a baseline to test alternatives against. But again, I wouldn't care for speed at all - I think that's a very minor thing to worry about because most of the time should be spent cutting and not moving around, and you don't cut at those speeds. Depending on the software you should be able to configure tool paths to have the shortest distance when moving between different paths. Perhaps you intend to put a toolchanger at one end of the table? That's the only reason I see to focus that much on speed given the compromises. I think if that's the case you might want to consider having the tool changer on the "across" axis arm there. Perhaps it's also added weight and inertia where you might not want it, but that's something you can make entirely out of plastic and aluminum so I personally would be willing to make that trade. Good video though, appreciate it.
Hi TrisT, thanks for your thoughtful comment. I hadn't actually considered the centrifugal effect on the oil - really interesting - perhaps a reason to use slightly sticker grease than I otherwise would. I know others use the system (rotating nut) with good results. Thanks again, Bongo.
You'd have to explain to me how a toothed belt "loses a few steps in precision when switching direction". A properly tensioned and sized belt will be as exact as a gear drive. Belts wear over time but are in comparison, easy and cheap to replace.
@@timb1179 I left this comment more than a year ago, don't exactly remember what I was thinking tbh. But those are high speeds and so I was likely referring to either lax tensioning, or just the stretch in the belt itself, amplified by the fair bit of inertia. I'll concede it doesn't seem to make a lot of sense.
@@Tristoo Fair enough. I am in the process of designing CnC myself and I am planning to use a driven ball nut design also. However I am buying pre-made units from a Chinese supplier. I went with a 32mm x 10mm pitch screw and driven nut.
Hi there just wondering if you have an email or if you would do a short whats app, I am based in Ireland, I am currently speaking with bstlinear about some ball screws and linear rails would lie to get your opinion and advise.
You showed us about 20% of the math's involved in the inertia equation So the nut is at a bigger radius than the screw so that needs to be put in the calculation you are also now carrying the mass of the motor around on the gantry with would be stationary if you have the screw turning you are also now turning those 2 or 4 gears to do the belt ... at what RPM was the ball screw sounding unpleasant .
Why invent the wheel again? Just go rack and pinion or just do it like it's normally done and rotate the shaft...When you rotate the shaft you can eliminate exces failure points...
Hey Jon, It's actually the 'ball nut' itself that's the limiting factor, rather than the opposing angular contact bearings. At 32mm dia. cheap ballscrews tend to get cluncky after about 1000rpm...
I'm a CNC designer, and I can tell you that make the nut rotates is the right choise on a long and/or fast screw. But you will need to constant lubricate the nut manualy. Normaly when we use a rotate nut, we make a much more complicated assembly so we can make a automatic lubrication. Very nice work. Cheers from Brazil.
@zomgthisisawesomelol It's a good option on long and fast travels, but you'll need a reductor (with low backlash) on the servo to meintain the milling force. For large portal machine you ussualy need 2 servos, one on each side. And the mounting system needs to keep a pressure from the pinnon on the rack, the eliminates the backlash from de rack. Cheers
Lubrication isn't that hard just take the seals out of the ball nuts inside the bearing housing and have an auto oiler drip oil onto the screw. As for the screw being a bit bent they are pretty easy to straighten, all you need is a set of v blocks a dial gage and a long pry bar or hydraulic press, you just rotate the screw on the blocks find the high spot and bend it back with the press/bar rotate again and measure and keep repeating until you get it straight. Just make sure you use a block of wood or rubber between the bending implement and the screw so you don't mess up the screw surface. As for the pitch you ideally want 10-15mm on a gantry that size. Looking at those angular contact bearings being grease sealed they probably top out at 6-8k rpm I'd guess. You probably don't want to run the ball nut much more than 3k rpm it really depends on the ball size and recirculation path though, longevity will be hugely dependent on adequate lubrication and keeping dirt out, there is a company out in California Called balltec that can supply balls in any size you like down to the micron if you need them, you could potentially polish the internals of the nut then reball to get them running more smoothly, I have rebuilt a lot of ball screws, many people treat it like a secret art but to be honest as long as you don't need to have them reground you can do everything yourself. The main reason for a stationary nut is so you can flood oil lubricate it to push contamination out! That alone leads to improved longevity even if the seals are non optimal, typically most vmc's pump a shot of oil into the screws and linear ways around every 8mins then they have an oil skimmer in the coolant tank to remove that non soluble tramp oil from the coolant.
Please make videos about ever aspect of cnc design. Professionals are not on UA-cam; ending with lot of miss information on youtube. Can you recommend a book about cnc design.
Very interesting way to turn the problem on its head. Makes me think of a helical rack driven with a gearbox. Except unlike the geared helical rack you’re not having to deal with (much) backlash.
Hey thanks man 😊. Was just watching your bow vid. Inspiring stuff my friend! Thanks for watching
Dont get why even spend money at ballscrews. Just use a normal 20mm threaded rod with 1mm pitch. Screw on it 3 or 4 or 5 screw nut and tight them together with hand. Then weld all together ,put in lathe and make it nice .The result is 0.000 flashback and always accurate. Will bend a minimal if so long threaded rod, but is doesnt matter while you turning the nuts and using the linear rail.
@@laszloszell8753 wear is the only concern! Over time the accuracy of the machine will reduce especially for large machines
You make something that involves real engineering, moving the nut involves less moment of inertia, congratulations youre a genious !
And those castings….
Hello friend, I want to thank and congratulate you for sharing, I enjoyed this system immensely, it's amazing
Was thinkning of upgrading the work 5x10 router to this so it can handle steel, now makes me think that I'm gonna have to put in a lot of care into designing the rotating nut mechanism and finding good screws... great vid!
The rotational inertia should not be limited to the mass difference of the nut vs screw, but how much of that mass is distributed away from the center of rotation- I suspect that the rotational inertia is not as much of a design constraint as you may think- the inertia of the gantry will obviously be more of a factor, so being able to put the motors on the table rather than gantry would be nice. And as another person posted, spinning the ball nut will complicate the oiling system. But there is never a right way or wrong way- very nice work and good job thinking outside the box!
Nice. I've never thought of solving the whipping problem that way. I built a 3m x 1.6m cnc router 12 years ago and it's got 3m long lead screws with recirculating nuts like the ones you had on yours. I do get whip quite badly when the gantry is at either end and it's moving at full tilt. Full tilt is about 100mm per second. However the pitch for mine is 10mm so it's not as bad as yours would have been. I've learned to live with it. For the person that suggested using a chain or a belt, the reason that I went with lead screws is that it's so much more accurate. I can make watch parts ....or I can cut out kitchen units. Anyway, thanks for sharing.
they put a little stretch on ball screws in standard cncs. but the castings are larger and more rigid also.
Driven nut design is a popular and proven way of removing whiplash.
You can use way smaller ballscrews since your spinning the nut. I designed a similar machine but ran the ball screw though the center of the steppers I'm using.
I've never seen that idea before, this is awesome! Also if you didn't show you welding, i would have never guessed that you built the pulley shaft from scratch
Thanks Eric :)
My 0.5" X axis acme threaded rod is over 6 feet long and I had never thought of rotating the lead nut rather than the rod (rod is hooked to stepper currently). I think between whip and rotating mass reduction, you have a great idea! But reading comments below, using a stationary chain and driving sprocket is a great idea too.
Hey Christian, thanks for this. The chain and sprocket would work well for low precision high-speed applications. Not so good if you want to mill aluminium to a decent accuracy... Too much backlash and/or inaccuracy in the chain.
@@FloweringElbow4SuperNerds Good input. I will go for the rotating lead nut; simple to do in my case. simplecncrouter.blogspot.com/2010_04_11_archive.html
I really like your channel. You're very earnest on camera and can tell you're excited to learn from the process.
Thank you for this video, it was soo needed on youtube!
( I'm planning the same rotating ballnut system )
Good luck!
Really a nice concept, I also made cncs in which the screw rotates as long as 12 feets. I will try this in my next cnc.👍❤️
Отлично получилось!!!👍 Но шапка понравилась больше всего😉 весёлая!
i wish i have a neighbour like you to work with.
It's good way to get around whipping problem when using long lengths
nicolas correa uses the (spindle nut) setup on their large cnc machines as well!! there is 1 at my work with that setup and it works well!
Nice design !! Good idea ! Thanks. Not a huge fan of the gears, you could find a stepper motor and put the ballscrew in it.
Inductive proximity sensors are very repeatable, dirt, wear and corrosion resistant and not particularly expensive. Dust will likely cause a pogo pin to jam over time.
The very best combination is inductive sensor then encoder zero index pulse. I've done inductive only and zero index on 2 different machines, zero index is dead accurate and I know I can resume a job the following day with the exact same homing position.
@@SuperYellowsubmarin Hybrid steppers will give you that but unless you're getting missed steps then once homed then any machine is going to be that repeatable, I only have to re-home my machine after a power down or crash. Inductive sensors are repeatable to a level far higher than most people need
@@mfx1 I mean encoder index as a means of homing ! Centroid Acorn allows that. Of course for a router it is not needed, bot for a mill or lathe I actually noticed the difference.
It's nice to see the inner workings. Langmuir systems is coming out with a new 4x8 plasma table that uses this method for driving the gantry. First time I saw it I thought it would be a genius way to handle screw whip. Anyways good luck with the build, it looks awesome so far.
Shop Sabre has been using rotating ball nuts on their larger machines for many years. Its a great solution.
How about multi start thread? If you can find/make those, that should make that machine fly. Good job for building such a badass machine.
First thing I thought about when I saw your layout is the issue you are going to have with chips getting into the threads of the shaft. Need to look at placement of the screw on the underside or at least on the outer edges where you can put some shielding over them. Also, definitely new two shafts with that placement or one underneath centered. Keep up the good work
Hi Chris. Thanks for sharing your thoughts. There are two ballscrews, one each side... They are situated on the outer edges, but you are right, we'll need to setup some kind of shielding.
@@FloweringElbow4SuperNerds Bellows are wonderful things. Just had to replace a set on the Bystronic laser cutters at work. They do a fantastic job
When the boss asking, what you guys was doing since the morning? Wait boss, I show you, check out this video 🤣
This is excellent work. I love the rotating ball screw. Can you just fill that housing with grease to prolong the ball screw life?
Hey Joel, thanks 😊
Recon we could fill the housing with grease. Still working on how to seal the dust and chips out of the screws though, so until then, I think it might attract contaminants like a magnet!
Very nice work! It looks like the problems are not coming from the rotating nut arrangement (which to me is genius, good choice on your part!) but from the too-fine pitch...
Eros - that's exactly right!
@@FloweringElbow4SuperNerds How find it in india
You gotta ❤️ the hat !
Nice work on the driven nut I am working on a 3 meter screw. Driven nut Ballscrews are hard to find
Hi, I really enjoyed your video series in making your industrial size CNC. I am wondering if you have the video of the process in making the rotating ball nut. I watch what you had posted carefully and wanted to know how you got the gear installed. Thanks... BZ
I love your hat.
If you are using actual servos many have a hard stop homing/stop by count so you might want to check more advanced settings...I hard stop and home by count.
Very Nice 👍 Good Idiya . I used 10 Feet 16 mm Ball screw. But wobling. I well try this.
Great work this is watch I was looking for diy example of this for a large format cnc!
Hi Derek, good to hear :) More CNC vids to follow...
Thank you very much for this, i really like this idea - gonna look at something like it for my plasma table build instead of fighting rack and pinions. :) (but using 10-15mm p itched screws) :P
Is a Smart idea, this is a way to move less weigt, if the screw it keep static the motor work less.
Great design. I suppose the only disadvantages I can think of are that the motor has to sit on the X gantry, which increases it's inertia (although probably not significantly at all) and that the additional wiring for these motors that needs to come off of the moving parts, meaning more complexity and more vulnerable.
On the other side you do not have an inertia problem with rotating the ever longer screw. THAT likely is a lot more intertie than the ADDED weight of the motor starting with a specific length.
@@ThomasTomiczek I think it would depend on the specific configuration. I think its also the kind of problem that you dont know the ideal solution to unless you build it and test it.
Regardless, I doubt that the inertia is the limiting factor, especially because CNC machines generally run pretty slow. Mechanical bandwidth is probably more than high enough.
@@tHaH4x0r There is more than just metalworking cnc. A woodworking router cannot be called a slow cnc. The "holzher ecomaster 7113" cnc also has the design of a ball screw. Nevertheless, the speed is 60 m / min at high speed (g00), 3500 mm in the X-axis direction. True, later models were already made with a "rack'n'pinion" design because they are cheaper. . .. At least I think so.
Nice video, thanks :)
Was the ballnut originally designed to be driven or did you modify it? If you did, then how did you affix the gear teeth?
I built a shaft that holds the pully. You should be able to see in the vid...
Without that cap is impossible to build such mashine ...I want it !!! 😭
nice work
Hey thanks man :D
call me crazy, but what about a combination of both?
nut rotating twice the speed of screw makes for a 10mm pitch. much faster rapids.
conversely, driving it at half speed gives 2.5 or something. higher resolution.
may need some fancy software to morph between high resolution to coarse as it feeds and rapids... then again, its just differentials... easy to mentally picture, hard to write as an equation... :D
Nice hat 😁
Thanks for making videos on the rotating ball nut design. There’s not much info about it.
Secondly if the motor is rotating too fast change the pulley confirguration to get a higher gear ratio. Less rpm for the motor and more rpm for the ball nut
Hey Jamie, thanks for watching ;) The real problem I have is too much rpm at the ball nut. In order to get good speed the ballnut's got to turn too fast, and it wont last... I have some new ballscrews with a larger pitch now ;)
I have a question about this. Since there are limitations with the top speed of the shaft rotating when very long, and there are also limitations with how fast you can spin the ball nut. What if motion happened by rotation the shaft at some arbitrary rpm and then you also rotate the ball nut? So speed would then be multiplicative.
Yes it is possible, but very complicated. The relative speed is the important limitation (how fast the ball bearings in the nut are moving round their helical path against the screw...
Hi friend... I don't know if you read these comments... I actually enjoyed your build.... although, I am a designer, and I want to put this into CAD... specifically SolidWorks V2022. And I have a few changes I want to make. I would enjoy to benefit from some of your experience... such as what size your ball screw is, and where you got your parts. What size Servo Motors, and where did you get your controllers and other electronics. I am a mechanical designer, not an electronics guy. That is why I am asking for such information. I hope you can help me. Have you put together a materials list and where to get the materials? Also some tips about setting up and software you are using for your setup. I do have access to CNC Mill, and a nice lathe and surface grinder and other equipment so I can do this project. So... if you would be inclined to assist me I would really appreciate it.
If your willing to assist me... I am willing to provide you with my designs, and also the plans I make. P.S. setting up the electronics is going to be an issue for me. I can put together components to build computers... so I do have some experience with playing with electronics. But I have never worked with controllers and servos.
I can be reached at... I am Leslieallen inventorshaven@proton.me Please contact me. Friend... if you have Signal Messenger it would be a big help, that way we can easily communicate using video or messaging. I prefer video actually. Anyway, write to us if your willing. Thank you. I will post this on each of your pages... so hopefully you will see it.
Appreciate your work, I m trying to build my own heavy weight CNC, can you tell me about servo motors and drivers ?
Hey jazzyyy. Thanks man, I'll try and do a video on the servo motor setup soon...
@@FloweringElbow4SuperNerds thanks bro, please make a video in detail 👍🏾
Trying to think outside the box, could you put a motor on the ball screw as well as the ball nut? Leave what you have for " fine control", but also drive the ball screw in the opposite direction for fast movement? This might be overly complicated, and not worth the effort, but might save you the money of new ball screws.
Hey Shale. That's a really interesting idea. The problem I have is that this would not reduce the speed the balls in the ball nut travel (instead they would increase proportionally with the added speed of the ball screw.
@@FloweringElbow4SuperNerds
Yes it would. I did not fully know if your issue was the speed of the ball bearings themselves, or you might be having an issue with how fast the ball nut its self was spinning causing the issue. By splitting the driving force you wouldn't have to drive the ball nut it's self as fast. Yes the speed if the ball bearings would be the same, but may reduce the issue. Again, a pretty overly complicated solution, but I don't know how expensive those ball screws were, and may be an idea if they were expensive. Of course the idea has its flaws and may not be a good one
Great information's,,thanks,,how is durability and accuracy,,,i think its better than rack and pinion system....
Hey, I'll let you know when I get the new ones installed and run it a bit ;)
But yeah in general, I think they can be a touch better for those things, but all depends on the details...
@@FloweringElbow4SuperNerds yes,,,,thank you,,
@@FloweringElbow4SuperNerds pleas can u tell me this bearing modle number
@@yantechlankamachine8767 Which bearing are you talking about?
@@FloweringElbow4SuperNerds fixing with ball nut..i think tat angular contact..also pleas give me details of belt and pully for it..thanks
Great stuff!😁 I'm considering a rotating ball nut design on a 1.5m long axis to avoid screw whip. One thing I can't figure out is how to lubricate the nut; as it's spinning I can't just add a lubrication line! How are you doing this on your build?
A the moment I use the oil nipple port on the ballnut when its stationary. That or I just put some oil on the ball screw itself. Crude, yes, but it works.
@@FloweringElbow4SuperNerds Thank you! That's pretty much what I'm doing with my current router; built before I knew about building in lubrication systems...!😁
Hey where did you get your ball screws
When you were calculating for the inertia you didn't account for the mass of the carriage itself since it is attached to the nut assembly, not sure if this is thr best choice. Will be interesting to see how you solve this.
Why? When turning the whole ball screw then no carriage weight?
If I understand correctly what you are saying, the mass of the carriage remains the same whether you are rotating the screw, or the nut. The inertia involved here is the comparison between spinning the nut, or spinning the screw... the balance of the inertia remains constant either way.
Is this as an experiment or trying to getting it to work? A Rack and Pinion system would work better for movement wouldn't it?
Hey Janak. It's both an experiment and I'm trying to get it to work! Thanks for watching, Bongo.
Kindly show me how to fix Ball screw, any loosing screw may slip and server motor will move to wrong position. i need to learn about it
hola. Cometí el mismo error en mi CNC, si ves videos de maquinas profesionales (Biesse) observaras husillos de paso 50mm/rev. Si piensas actualizar a motores de bucle cerrado te daras cuenta que en altas rpm pierden hasta el 80% del torque. Hay que hacer que los motores trabajen a baja rev donde su torque es mayor. Tambien ayuda usar tpreload double ball nut.
Hola, gracias por tu comentario. Es bueno saber que no soy el único ;)
Estoy usando servos AC de segunda mano, que parecen estar funcionando bien hasta ahora. Obtuve tornillos de paso de 20 mm al final :)
You can see it in action here: ua-cam.com/video/obR-l7qVhQY/v-deo.html
Excellent setup. I read that you increased your ballscrew pitch to 20 mm. Was this because you were on the limit of your servo motor RPMs? I'm on the final stage of a CNC router too and I'm considering purchasing 10 mm pitch ballscrews and using rotating nut too. Ideally I should get 20000 mm/min easily with the servos but I would love to know your experience rotating the nut that fast. Thank you :).
Hey Leonardo, thanks for getting in touch. Yes I now have 20mm pitch TBI screws (see them in the latest vid). The reason was that I the ballnut has a realistic max speed of about 1000rpm on these cheap screws... More than that and they make a lot of noise - and wont last long! The 20mm ones I got have a different kind of recirculation path for the balls and are designed for higher speeds - but still have a max of about 2500rpm (but travel should be plenty fast enough a reasonable amount below that speed with 20mm pitch).
Are you planning rotating ball nuts or to spin the screw?
@@FloweringElbow4SuperNerds I'm planning to use rotating ball nuts as you. I'm double checking now if the ones I'm going to use have a RPM limit close to 3000 rpm which could be a problem because I really need a 3000 mm/min speed to be comfortable.
@@leonardomarsaglia It depends a lot on the diameter of the screw I think. Mine are a bit large at 32mm - even ground expensive THK ones have a max speed less than 3000... Best check the documentation. What make is it?
@@FloweringElbow4SuperNerds Sorry for the late reply. The brand is PMI. But I could find some 25 mm diameter 25 mm pitch screws and now I'm finishing the mounting setup for those. Thank you :)
Where did you get the ballscrew ? I thought about either buying a hiwin or tdk ones..
the one in the video is a cheap one. See video description...
What do you plan to cut with this machine?
wood, aluminium and plastic.
Design Option 1: Steel Rod Rotating at 100 revolutions per second
Moment of Inertia of the rod (I): 0.40625 kg·m^2
Angular Velocity (ω): 628.32 rad/s
Rotational Kinetic Energy (E): (1/2) * I * ω^2
Result: The rotational kinetic energy is approximately 80,562.5 joules (J).
Design Option 2: 10 kg Object Moving at 70 cm/s
Mass of the object (m): 10 kg
Velocity of the object (v): 0.7 m/s
Kinetic Energy (E): (1/2) * m * v^2
Result: The kinetic energy is approximately 0.245 joules (J).
Conclusion:
Comparing these two design options reveals a significant difference in terms of kinetic energy. The steel rod rotating at such a high speed (100 revolutions per second) possesses substantial rotational kinetic energy, approximately 80,562.5 joules, due to its moment of inertia and angular velocity. On the other hand, the 10 kg object moving at 70 cm/s has much lower kinetic energy, around 0.245 joules, due to its lower linear velocity. This difference in kinetic energy is crucial when considering the design and application of these systems.
how did you fix the ends of the lead screw?
Kind of show this in this video: ua-cam.com/video/k47LBPEemvU/v-deo.html
Essentially with four set screws per side for alignment, and threaded at the end to tension the screw.
Number 217 suscriptor. 😁
ball screw needs to be in the middle. or there needs to be 2 of them, one on each side. ive got a professional built laser table and a home built plasma. been through this a time or two.
this is fun but insane, rather use a rack and pinion, or just a stationary chain snaking through a u-bend over a driven sprocket in place of your rotating nut.
I've seen rotating nut setups before but never used one, they always looked the business. I once read that you should tension the ball screws too on a setup like that. What's puzzling me is what pulley are you using on the pipe? I thought it looked like a taper locked pulley but after closer inspection I noticed it isn't. I can't figure out how the (driven) pulley is fastened to the pipe.
Hey Paul, good question. The pulley there is actually bonged on with locktite. It works to hold it, but a right pain to remove the pulley if you want to change it - I know!
@@FloweringElbow4SuperNerds Thanks for replying... That Loctite is seriously strong, to remove it we used to get a blow torch on it. I might be wrong though. Did you buy a regular pulley and turn the middle out on a lathe or did it come with that bore? I'm curious to know as I've been after making a rotating nut setup for a while.
@@paulpavlou515 yeah, I used a hot air gut and a lot of patience when I removed...
Yes I did turn out the middle on the lathe...
Bom trabalho
I see a lot of videos where people want fast CNC. Personally I'd rather have slow and precise vs fast and sloppy.
Sometimes speed is a requirement, depends on the material being cut. There's no reason you can't have speed and accuracy
Hi can anyone tell me if is possible to have a play in the nut if I rotate the screw manually not one full turn just a little bit let's say 20 percent and the nut won't move is that normal ?and what could be the cause tnx
Where to buy that rotation ball but? And what size?
Hey friend. Many many MANY options and design choices in that there question.
At what rpm it is rotating ?
about 2k is the max shown. That's probably quite a bit more than it should be for a ballscrew of this diameter and quality.
Такой тоже собираю .
hi i have a cnc router i need to change the motors of a 3 meter x 2 meter router, and i was considering buying absolute encoder servo motors with standalone gsk controller, can someone guide me, is this feasible? or only servomotors with incremental encoder would be enough.
Hey there, this is outside my experience so I can't offer much advice I'm afraid, hopefully someone might chime in... Good luck.
Hello Hanz, I would go with 17 bit absolute encoders all the way. You can tune your servos perfectly with that. If you are planning to purchase chinese servos, you'll find there's not much difference in price between 2500 PPR and 17 bit absolute. I'm using delta 17 bit absolute servos on a lathe and they are wokring flawlessly.
Having a rotating ballnut might not be much better than a static one. The inertia is rotational inertia which means that it depends on the mass distance from the radius, for your monster balscrew it will help but for smaller screws the difference will be less. Also the whole machine gets more inertia as the motors travel with the gantry. Lastly you also have to have bigger cable chains for all the extra wiring.
That being said, for chinese ballscrews which might not be straight this system minimizes vibrations and the risk of the ballscrew throwing around.
Actually the rotational inertia of those long ball screws will be ~5x the inertia of the gantry. Rotating the nut is the way to go.
А что будет хуже если покрутить не гайку а винт?
Инерция будет больше, если винт повернуть. Это затрудняет ускорение и замедление серводвигателя.
@@FloweringElbow4SuperNerds в моем станке супорт весит тонну...он останавливается на скорости 10 метров как вкопаный...шо вы выдумываете?
My brain isn't a fan of that. Most of the rotational inertia would be very near the centerline, so I don't know how much of those 15kg actually matter - and I'd wager the inertia of the 2 other axis probably dwarfs it. Also that just seems like a great way to centrifuge oil away from the ball bearings (especially where they meet the screw, which is most critical). Mostly though, that just sounded terrible. I'm also not sure that belt bodes well for precision since the encoder is in the motor - I'd wager you're losing a few steps of precision when switching directions. If it were me I'd probably just rotate the screws, at least until I got a baseline to test alternatives against.
But again, I wouldn't care for speed at all - I think that's a very minor thing to worry about because most of the time should be spent cutting and not moving around, and you don't cut at those speeds. Depending on the software you should be able to configure tool paths to have the shortest distance when moving between different paths. Perhaps you intend to put a toolchanger at one end of the table? That's the only reason I see to focus that much on speed given the compromises. I think if that's the case you might want to consider having the tool changer on the "across" axis arm there. Perhaps it's also added weight and inertia where you might not want it, but that's something you can make entirely out of plastic and aluminum so I personally would be willing to make that trade.
Good video though, appreciate it.
Hi TrisT, thanks for your thoughtful comment. I hadn't actually considered the centrifugal effect on the oil - really interesting - perhaps a reason to use slightly sticker grease than I otherwise would. I know others use the system (rotating nut) with good results.
Thanks again, Bongo.
You'd have to explain to me how a toothed belt "loses a few steps in precision when switching direction". A properly tensioned and sized belt will be as exact as a gear drive. Belts wear over time but are in comparison, easy and cheap to replace.
@@timb1179 I left this comment more than a year ago, don't exactly remember what I was thinking tbh. But those are high speeds and so I was likely referring to either lax tensioning, or just the stretch in the belt itself, amplified by the fair bit of inertia. I'll concede it doesn't seem to make a lot of sense.
@@Tristoo Fair enough. I am in the process of designing CnC myself and I am planning to use a driven ball nut design also. However I am buying pre-made units from a Chinese supplier. I went with a 32mm x 10mm pitch screw and driven nut.
Whenever you got the chance please do a video about that freaking object that sits on top of your head .
haha.
Wooow
What? But the portal is many times heavier than the screw.
What do you mean by the portal? The gantry?
@@FloweringElbow4SuperNerds Yes. We call it portal (портал) in Ukraine/Russia.
@@humanspirit3432 ah ha, cool. Well, it has to do with the spinning inertia... Makes quite a bit more different than the mass of the portal...
👍
Hi there just wondering if you have an email or if you would do a short whats app, I am based in Ireland, I am currently speaking with bstlinear about some ball screws and linear rails would lie to get your opinion and advise.
Might be too late now sorry, but yeah, drop me an email. Go to the 'about page' of this channel to find the address.
You showed us about 20% of the math's involved in the inertia equation So the nut is at a bigger radius than the screw so that needs to be put in the calculation you are also now carrying the mass of the motor around on the gantry with would be stationary if you have the screw turning you are also now turning those 2 or 4 gears to do the belt ... at what RPM was the ball screw sounding unpleasant .
Hanzhen harmonic gear , over 30 years experience , industrial robot arm gear
Thanks
😂😂👍👍💪💪😎😎⭐️⭐️⭐️⭐️⭐️
Why invent the wheel again? Just go rack and pinion or just do it like it's normally done and rotate the shaft...When you rotate the shaft you can eliminate exces failure points...
just get better bearings? 1600RPM isn't really that much for a small bearing.
Hey Jon, It's actually the 'ball nut' itself that's the limiting factor, rather than the opposing angular contact bearings. At 32mm dia. cheap ballscrews tend to get cluncky after about 1000rpm...
@@FloweringElbow4SuperNerds sorry I must have smoked something that day. I guess that's the drawback of low pitch.
@@jonjon3829 yeah, right on. Thats why I changed to 20mm :)
Just a small note and not at all a criticism of your build or achievement but that isn't a large cnc, not even close. Its small to mid size.
lol. It's mig compared to me! Point taken though :)
@@FloweringElbow4SuperNerds Fair enough ;)