-
3 Attachment(s)
Small Concrete CNC Mill
Hi everyone,
I've been thinking about building a CNC mill or a router for several years. These are some of my thoughts:
Objectives/considerations:
-Main goal is learning about CNC in general and being able to machine plastics, wood and soft metals like aluminum.
-I don't have much space nor budget, so that has to be taken into account. I'm happy with a 200x200mm work area.
-Cost must be kept low.
-I don't have particular tolerance requirements. I think I'll be happy with +- 0.1 mm for most things I do.
Tools available:
-I have a small lathe, drill press, angle grinder and stick welder.
The plan:
I was thinking of making an L-shapped reinforced concrete base, from which threaded rod would protrude on four places on the base and on other four on the column. Those would be used to attach 20 mm thick aluminum plates, making their position slightly adjustable so they can be aligned, particularly if concrete shrinks or what not. Those and other necessary plates I would have cut to size, and do the drilling and tapping myself.
On those plates I would mount some SBR20 rails and RM1605 ballscrews as shown on the pictures below. Three nema 23 steppers would run the show, and the spindle could be something like the ubiquitous RT0700C makita router which seems a good compromise for my situation
I would be buying one of those kits on ebay providing both the rails and ballscrews for three axis, and cutting those to length. Seems to be the most cost-effective option. For shortening the ballscrews, I think the easiest method would be to cut them with an angle grinder, drill and bore a hole on the lathe (as they are case hardened and the inside is soft), loctite a cilindrical piece in place and machine it concentric.
Regarding electronics, I would be using and arduino and GRBL for starters, as that's what I'm most familiar with.
I still have quite a bit of work to do on the design, but I'll be glad to receive any feedback regarding this approach.
Juan
https://www.cnczone.com/forums/attac...d=457412&stc=1
https://www.cnczone.com/forums/attac...d=457414&stc=1
https://www.cnczone.com/forums/attac...d=457416&stc=1
-
Re: Small Concrete CNC Mill
Round rails (SBR) are not suitable for a metal mill. Don't waste your effort.
Don't limit yourself to available "kits". Any decent seller will customise and cut to length / end machine (for cheaper and easier than you can). If they won't then they are just a drop shipper and not best to buy from.
I use Fred at BST Automation on AliExpress. I've done multiple orders and always happy.
-
Re: Small Concrete CNC Mill
Quote:
Round rails (SBR) are not suitable for a metal mill. Don't waste your effort.
Don't limit yourself to available "kits". Any decent seller will customise and cut to length / end machine (for cheaper and easier than you can). If they won't then they are just a drop shipper and not best to buy from.
I use Fred at BST Automation on AliExpress. I've done multiple orders and always happy.
Thanks for the quick answer. I suppose that suitability would deppend on the speeds and feeds, depth of cut and so forth. I'll have to run the numbers, even if the rated load for SBR rails is substantially smaller than that of profiled rails, they may suffice. I originally chose them as they are easier to mount and more forgiving with alignment. I could mount them together with the ballscrew bearing blocks on a single aluminum plate, which I would expect to be reasonably flat. I have no means to propperly mount and align profiled linear rails. I could try to figure something out though. BST Automation seems to be widely recommended, so I may purchase from them.
What about the general disposition of the machine? Are aluminum plates held to the concrete base with threaded rod going to be rigid enough?
-
Re: Small Concrete CNC Mill
Em..
upgrade your expectations and your budget.
A mill requires by definition the ability to cut steel with a cutter that is unsupported.
And that has minimal deflection / chatter at typical working loads.
0.1 mm is totally useless.
Milling machines need to aim for sub 0.001 mm theoretical, low, to approach 0.01 mm real world.
Look at existing watchmakers mills for inspiration and scale from them.
You are never going to do lighter than say HAAS for a particular work envelope.
Scale from that.
-
3 Attachment(s)
Re: Small Concrete CNC Mill
Quote:
Em..
upgrade your expectations and your budget.
A mill requires by definition the ability to cut steel with a cutter that is unsupported.
And that has minimal deflection / chatter at typical working loads.
0.1 mm is totally useless.
Milling machines need to aim for sub 0.001 mm theoretical, low, to approach 0.01 mm real world.
Look at existing watchmakers mills for inspiration and scale from them.
You are never going to do lighter than say HAAS for a particular work envelope.
Scale from that.
First one is easy, second one not so much...:)
So maybe I should stop calling it a mill, milling steel is not on my requirements nor my capabilities.
I know 0.1 mm is almost woodworking territory. What I'm trying to say is that I don't have particular tolerances in mind. The main goal is learning, and to end up with something I can use for my projects.
Regarding rails, I think I'm going with HGR15 profiled rails, and I'll buy those and the ballscrews in the lengths I need. For the price difference it's not worth it to cut them myself.
I've redesigned the frame, I think I'm doing away with the concrete. I just can't move around a 150 Kg concrete frame. So my new design uses 100 mm steel square tubing for the frame as shown in the images. Tubes will be screwed together and then welded. I'll fill it with sand or even epoxy granite if I see it appropiate. This will make the build more manageable.
I've done some calculations and FEA simulations and for a 1000 N load on the tool (which should be plenty more than what the spindle would windstand), maximun deflection stays within 0.02 mm for 5 mm wall tubing. On resonance and such I can't really say anything, but the sand should aid with dampening. I suppose bearings would be resposible for most of the deflection anyways.
10 mm steel plates will be screwed on to the frame to provide a flat surface, and some rectangualr sections are to serve as support for the rails.
So my plan now is to order rails, ballscrews, motors etc, and visit some of the industrial suppliers around to get an idea of what is available. I'll try to get some cold rolled sections so they are reasonably flat to mount the rails to. It seems advisable to get a scrap steel plate and fill it with tapped holes, so I become familiar with the process.
I'm happy to hear any recomendations. To provide some size reference, the machine fits in a 700 mm side cube, and the work table is 150*300 mm usable area.
Juan
https://www.cnczone.com/forums/attac...d=457634&stc=1
https://www.cnczone.com/forums/attac...d=457636&stc=1
https://www.cnczone.com/forums/attac...d=457638&stc=1
-
Re: Small Concrete CNC Mill
Hi Juan!
I literally made an account 5 minutes ago after spending a while lurking on the site so hopefully I'm not way off base here breaking any rules or anything.
It seems like you moved away from a concrete design based on your latest posted pics which is good, the concrete is going to move over time with different humidity levels and actually is not nearly as vibration damping or rigid as you'd think. There's a good video series by This Old Tony on youtube making a cnc router which sounds a lot closer to your expectations for actual machining capabilities (wood and plastic easy, aluminum without too much trouble). One thing though, when packing the tube stock with sand you need to ensure its dry and compacted. A vibrating table helps a ton as you see in his video, but you don't need anything too fancy, even like an orbital sander attached to your piece would help vibrate and compact it, or if you're really desperate a soft headed hammer and a whole lot of whacking would do the trick hahahaha.
By the way, that tube stock it looks like you're using is probably not going to be very accurate, and so bolting linear rails to it will make them inaccurate too. Do you have a method to increase the precision there? I heard of a guy using leveling epoxy which apparently levels to about 5 thou, which may work for your tolerances if not
Either way good luck with your project, hope it goes well!!
-
Re: Small Concrete CNC Mill
Hi Juan -
1) do not order anything until you have fully resolved the design. If you do order stuff this is your first major mistake.
2) Sand is a waste of effort. So is epoxy to some degree. Epoxy is expensive and you are better off using thicker tubing. Most first time builders use tube that is too thin and this contributes to vibration. Use thick tube say 9mm plus (see later do not use tube)
3) design a machine with as few parts as possible. By building up tubes with plates you introduce distortion, thin edges that vibrate, poor design. If you are to fabricate then use plate only. In this way you have control over geometry and local thickness. Everytime I have tried to design a router or mill using SHS or RHS it always goes AWOL due to std sizes. Design a plate steel machine. Will not need filling if stiff enough
4) I'd consider a no weld machine fully bolted up for first machine. Welding introduces all sorts of issues. There are plenty of good plate, bolted machines in aluminium or steel in the forum
5) Look up the Milli thread covers most of the territory you are going to need
6) good luck - cancal any orders that you can.... you need to develop the design much more before it gets into the wild...
7) your original idea with concrete is good, just need to learn alot more about it. Do not use portland cement but study up on UHPC or epoxy granite. Both are used to make very precise and large machines. Easy to cast at home has many possibilities...
Peter
-
Re: Small Concrete CNC Mill
Mook512:
I've watched those This Old Tony videos...multiple times. In fact it's what got me interested in the subject. Yes, it's pretty close to what i'm aiming for, but for size and cost :).
Will keep that in mind if I decide to use sand.
Regarding the tubes, yes I don't expect them to be flat or on dimension. My idea was to have a reasonably flat plate bolted to the tube, to which I could mount the rails, and behind which I could use some shims for alignment. I'll read about self-leveling epoxy, seems that it could be useful.
peteeng:
I haven't ordered anything. Too big of a purchase to make without careful thought. If anything I'll get some taps to practice tapping holes...
I've had thought about using plate, aluminum extrusions... An option would be to design an aluminum plate machine, aluminum plate beeing expected to be flat, and also easier to work with within my limited capabilities. Regarding welding and screws, I had thought of fixing the column with just screws so it could be slightly adjusted. But again, this design I've posted is what I've happened to be thinking of when I decided to post on the forum. I often change my mind about the design and I'm open suggestions. Luckily my requirements stay more or less constant.
I've gone through the Milli thread some time ago, lots of interesting configurations.
On the concrete design, the major disadvantages I see are weight and beeing difficult to modify. A screwed together machine can be disasembled, and if one part needs to be modified or re-done that can be done with relative ease. I'll investigate a bit more and see if I can come up with a new design.
Thanks,
Juan
-
Re: Small Concrete CNC Mill
Hi Juan - I have pretty much resolved Milli if you haven't been there in a while. Was a bit of a ride into new stuff. One thing to think about is to make it in timber. Its easy to change and work with and once you get where you want make the metal bits. There's a couple of builds here who build plywood machines, then use that to build the aluminium machine then that machine to build mills. Personally I'd stay away from construction extrusions, good for making stands but like std tubes have compromises that can't be fixed. An aluminium plate router/mill is very doable. Look up the Katran thread... Peter
addition - re modelling - I see you are not modelling the radius on the corners of the tubes. These must be included for a couple of reasons. 1) you may be tempted to place something at an edge that is in fact air 2) when it comes to welding the rads are an issue. They can create huge distortion if you weld along them 3) by modelling square you are increasing the inertia slightly so your deflection calcs will be better then they should be. Your FE says an applied load of 1000N and a deflection of 20um ie a static stiffness of 1000/20 = 50N/um. This is small VMC territory. You need to change the way you are applying load and holding the machine... that size machine and tubing is likely to <2N/um in practice... You need a machine at about 10N/um to do steel 1-2 will cut timber and plastic easily and may do aluminium in a struggle.... Keep at it - Peter
-
Re: Small Concrete CNC Mill
Juan ..
Your numbers are way off base.
The best manufacturers in the world are not getting 1/3 of the stiffness your numbers suggest.
A typical max load on a mill of 2 HP, bridgeport territory, is around 40 kgf.
400 N.
This will deflect about 0.1 mm on a bridgeport table, at the edge.
Go ahead.
Push on a table at a demo room, with a dti measuring the deflection.
2 fingers push of 10 kgf or so will deflect about 0.03 - 0.08 mm, give or take.
And the BP masses 2000 kg.
Gene Haas told me his VF6 series mills deflect about 0.01-- mm if you hit them with a hand, while they are milling.
Several microns.
The dimple is clear to see in the finished work.
At 6000 kg mass.
-
2 Attachment(s)
Re: Small Concrete CNC Mill
Quote:
Hi Juan - I have pretty much resolved Milli if you haven't been there in a while. Was a bit of a ride into new stuff. One thing to think about is to make it in timber. Its easy to change and work with and once you get where you want make the metal bits. There's a couple of builds here who build plywood machines,...
Yes, I've seen it. You have done some extensive design work. I'm sure it'll turn out great.
Regarding wood, I think a design based on plates lends itself well to wood prototypes. Or one could make the main components out of aluminum, and the more complex parts (read: not rectangular) in the spindle mount could be made out of wood and used to machine the definitive parts... I've seen the work of katran. Beautiful machine.
My CAD model was a bit rough and ready. Just to test out some ideas. That's why I didn't model the corner's radii.
I am a complete newbie on FEA if it isnt apparent yet. Those deflections were only of the bare frame, no table, no bearings, just the L shape and an additional section standing for the spindle holder. Base was fixed on a couple points. I'm using Fusion 360 BTW. I did calculate a load case by hand and results were near enough to those of the computer.
So real numbers, or those from better simulation taking other things into account would indeed be much larger.
Thanks for the stiffness numbers. I don't really have a sense as to what to expect and that really helps.
Quote:
Juan ..
Your numbers are way off base.
The best manufacturers in the world are not getting 1/3 of the stiffness your numbers suggest.
...
Thanks, it is a nice thing of online forums like this that one can get an idea of what things are really like.
I'm more of an electronics person. Now I am like someone in an electronics forum who asks how to charge a 46F capacitor with 15 KV...
----
I've done another mockup based on 20 mm aluminum plates. I have run another FEA on it, this time taking into account the table, rails and bearings. I have specified the bearings as being made out of aluminum, as I have read somewhere that that aproximated the actual behaviour of real bearings. Contact between plates is automatically specified as "bonded", but I can add screws and specify those as the contacts for more accurate results.
I have added a 400 N force on the spindle holder and fixed the table. Maximun deflection clocks in at about 0.065 mm, so that is 6.2 N/um. With a better placement of the load, and rising the spindle support to its upper limit, deflection would be higher. That seems more reasonable, although maybe still too optimistic. See attached.
So that could be a plausible approach, everything screwed together, already flat-ish aluminum plates.
Motor mounts and spindle holder could be made out of wood or 3d printed, then new version machined...
Juan
https://www.cnczone.com/forums/attac...d=457698&stc=1
https://www.cnczone.com/forums/attac...d=457696&stc=1
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan - Now your looking good. Rails should never be mounted on "air" so move the side plates inboard so they are under the rails. Use an alternating bolt pattern so the rail bolts thru to the webs. Similiarly with the spindle arm, put in a central web. 20mm is probably over kill for your machine spec, 16mm will be good if the details are right.... If the base and arm can be cut as an L it will be much stiffer.... So the big trick with this construction is having the plates cut (or milled) square and flat... then there is all the edge holes, so many holes & threads but then what else are you to do in the evenings? :) Keep at it Peter
If you upsize your rails the stack height delta will be less so the spacers will be less height. The cost delta is small overall. Rails are very important for local stiffness. You can remove the spacers by using the web edges for the rails see attached... the aim is to minimise the part count... Your config is traditional and solid. I have no doubt it would cut aluminium if made in good quality plywood. The Mikita is a bit fast for Al but consider making it a bit bigger in solid timber (screwed and glued, epoxy sealed) and then use that to make your alum version. This will get you thru the learning curve of the electronics, wiring, chip loads, cutting set ups, CAM and many more, the journey is still long.
-
Re: Small Concrete CNC Mill
Great, thanks.
I'll make a new design including your considerations. I'll also try to get hold of a catalog of aluminum profiles from a local supplier, maybe I could use some suitably sized profiles that can be just cut to length. I may have access to a CNC router at my university (though I've never used it, and I'll have to ask around), I could buy some aluminum plate and cut all pieces from it, even center drill hole locations. Sending it out to a local job-shop would be expensive...but worth checking. I didn't mount the rails as you've shown on the sketch because that would need machined edges, a saw cut finish would not suffice. But if I use suitable profiles or have the plates machined that would be no problem.
Being an engineering student evenings do fill up pretty quick :) I am on "holidays" at the moment. I hope to work on the design now, so any parts I have to order overseas arrive before summer.
Juan
-
Re: Small Concrete CNC Mill
Hi Juan - You can cut all non critical cuts with a good jig saw or radial saw. Then get the critical surfaces cut at a machinist maybe. If you use extrusions look at heavy channels they will be the best bet. But they are not square... Everything has limits. You need a CNC to build a CNC. Seriously consider a plywood first machine to make the next machine... just make it big enough to do the parts you want in Machine No2! Takes about a year to build No1 so don't panic yet...
https://www.cnc-holzfraese.de/
Peter
-
2 Attachment(s)
Re: Small Concrete CNC Mill
Quote:
Your config is traditional and solid. I have no doubt it would cut aluminium if made in good quality plywood. The Mikita is a bit fast for Al but consider making it a bit bigger in solid timber (screwed and glued, epoxy sealed) and then use that to make your alum version. This will get you thru the learning curve of the electronics, wiring, chip loads, cutting set ups, CAM and many more, the journey is still long.
Sorry I missed this part last time.
Quote:
Hi Juan - You can cut all non critical cuts with a good jig saw or radial saw. Then get the critical surfaces cut at a machinist maybe. If you use extrusions look at heavy channels they will be the best bet. But they are not square... Everything has limits. You need a CNC to build a CNC. Seriously consider a plywood first machine to make the next machine... just make it big enough to do the parts you want in Machine No2! Takes about a year to build No1 so don't panic yet...
https://www.cnc-holzfraese.de/
Peter
The thing is, if I'm able to make a machine out of plywood that can already cut aluminum good enough to make another machine, I would have already fulfilled my requirements and there would be no "need" for another machine :). Yet the primary goal is learning, and I'm sure I'll have plenty of fun with a plywood machine and getting the rest sorted out as you say. I could upgrade the frame to aluminum in the future if I saw the need, and keep the rails and such. I could even use the makita for its intended purpose and square up the wood pieces. Seems that a machine made out of plywood or even MDF would be stiff enough for my purposes.
I've made another design, see attached. The only thing is that attaching the rails straight to plywood doesn't seem right. I'll have to find suitable screws, but I guess it could be done. Alignment is another issue, although I suppose the wooden frame would flex before I risk damaging the bearings.
I see what you mean, but I don't really have the room or budget for two machines.
So we have gone from concrete, to steel, to aluminum and now wood. I guess the next material of choice would be wet noodles (just kidding).
Juan
https://www.cnczone.com/forums/attac...d=457770&stc=1
https://www.cnczone.com/forums/attac...d=457772&stc=1
-
4 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan,
MDF has a modulus of 4.5GPa while good plywood can be up to 21GPa. I get F22 which is 16GPa. See attached design guide... Re: screws I use standard metal thread screws. I use a tapered intermediate tap to make the thread. This provides a taper for the screw to tighten to. Once the parts are set up and straight I remove screws one by one and set them with PVA glue or epoxy. Epoxy is for permanent connections PVA if they need to come out in future. You can use wood screws I like the Kreg brand square drive. Fits into the counterbore of the rails as well. You can remove epoxied screws by placing a soldering iron on the screw to soften the epoxy then remove...
Here's an MDF machine base I'm building at the moment. From the base up its then stainless steel laser cut parts. The original intent was to use ply but melamine coated mdf doesn't need paint and its an experiment and half the price of ply. Mostly its pocket screwed, recommend that as well.
http://ewp.asn.au/wp-content/uploads...v5-FINAL-1.pdf
High Modulus noodles maybe something to explore. :) Peter
edit- The thing with wood is its not very dense (plywood 700kg/m3) so you could make this solid sections and it would not be very heavy. Instead of making all the webs etc I'd be laminating plywood into solid chunks then shaping them. Then there's no local deflections, perfect shear transfer and nothing to wobble. Timber is naturally damp so thats a win as well. Just seal everything well to slow down moisture ingree egress. Use PVA diluted 1:1 with water as a sanding sealer lacquer. Cheapest sealer there is since your $$$ short. Epoxy is the best but $$$. Peter
Since your thinking through materials here's something I wrote a while ago. Maybe laminated aluminium sheet and ply is the go...
-
2 Attachment(s)
Re: Small Concrete CNC Mill
Hi Peter,
Thanks for the tips on the screws and materials. And for the extensive plywood design guide.
Nice machine base.
On the laminated aluminum sheet and plywood, I think in my case it is easier to do just plywood, as it can be easily screwed and glued together. With the laminate it may be more difficult to join.
As I'll have to buy a decently sized plywood board (or two), making it solid should be no big deal. I've attached a new cross section, using 18 mm thick plywood, as that is what I can acquire locally. I've also added some motor mounts.
----
Regarding rails and ballscrews, I think 15 mm rails and 16 mm ballscrews should be fine. I don't think it's worth it to go into bearing life calculations and such as I'll be getting knockoffs made of who-knows-what. Judging by the listed static and dynamic load ratings, I suppose it should be ok.
On motors, I've played around with a spreadsheet from here: What size stepper motor do I-need., and it seems that something like this would do for my needs:
https://www.omc-stepperonline.com/ne...s30-3004s.html
I've considered the moving mass as if the frame was made out of aluminum and cutting feed of say 2000 mm/min, other parameters such as 80% efficiency for ballscrews and other dimensions as advised in said page.
Juan
P.D. Just in case, I'm based in Spain. I spent a year studying in Canada. No square socket screws or screwdrivers around here :). I've changed my location on my profile, but it seems it hasn't updated.
https://www.cnczone.com/forums/attac...d=457812&stc=1
https://www.cnczone.com/forums/attac...d=457814&stc=1
Edited: square socket instead of head.
-
2 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan - attached is the motion calcs for Milli- You can adapt to your machine. The screws are square sockets not square heads. Kreg have square drive and large head good for timber. There are torq heads as well. I dislike phillips or posidrive heads. If you glue your bits together you can remove screws after glue sets and reuse.
Using steppers means your torque drops off as you go faster so don't count on much torque past 500rpm. Usually design the structure fairly well so you know the masses you are trying to move around. Being small 12mm ballscrews would be fine... but check buckling and whirl (whip)...
Hiwin publish an excellent design guide for there stuff. speaking of bearings get medium or high preload cars if possible does make a difference.
something like below makes your first build easier instead of buying descrete drivers and BOB. This is a 5 axis board incase you want to add extra bits.
https://www.ebay.com/itm/5-Axis-CNC-...-/251315551321
Keep at it and its looking good. Peter
I use UCCNC but since $$$ are short go for linuxcnc its free and very powerful just need to know about linux which is very useful for you as a student into the future...
Make your table 2x thick. When your spindle plunges it will bend an 18mm ply. maybe the table is a candidate fore the AL/ply/Al
-
Re: Small Concrete CNC Mill
Hi Juan,
A note about steppers. They all look alike but there are some differences you need to know about.
https://www.omc-stepperonline.com/ne...m-4-wires.html
If you look at this one its a few cents more but its inductance is less. Inductance is an indication of its electrical efficiency in terms of how it generates back emf. The lower the inductance the less back emf it makes. The consequence of this is that the lower inductance motor will spin easier and faster at a given voltage. The usual fix for speed is higher voltage but high voltage systems cost more $$$. Now the one above is also rated at a higher current but if you use the driver I suggested you won't be able to take advantage of that extra because the driver is 3A. But thats not a bad thing. Comes down to $$$, torque required and get the lowest inductance motor you can get in the budget. Another thing is that a std motor has 200 steps per rev and there are 400 step per rev motors out there (there's a couple in stepperonline) and this doubles your accuracy for no extra cost. It may be worthwhile looking at N17 motors and 400 step in the stepperonline catalogue. Since your using a ballscrew you will have heaps of grunt from even small motors.. speed vs force always a trade off ... Peter
also saturate the ply edges with epoxy where you are mounting the rails this will harden the timber edge grain. Use a thin epoxy...
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Thanks for the calculations. I've done my best to adapt them to my case, se attached. Main asumptions are on the first page, an acceleration in the rage of 0.5 g seems to be typical and for speed, 4000 mm/min is pretty fast I think (I have a pen plotter here in my desk i can drive at any speed to get an idea).
I meant square sockets sorry. I'll try to get similar screws, but I guess philips head would do fine. It's just a matter of convenience.
I've seen documents by Hiwin. Looking great but hadn't had a chance to read much yet. I'll try to find a compromise option. Genuine Hiwing rails and cars are really expensive. I've seen some kits on Aliexpress that come with a couple 15 mm rails, four cars and a 16 mm ballscrew with its bearing blocks in the rage of 70-80 USD per axis including shipping. Maybe garbage, who knows. I'll have to keep looking.
I've seen this https://www.omc-stepperonline.com/cn...rice&order=ASC
kit on stepperonline, basically the cheapest 3 axis kit they have. It is what I have taken into account in the calculations. Motors seem kind of big, but may be appropiate. Having separate drivers appeals to me more, if something goes terribly wrong, I won't end up with a useless board.
I had thought of using grbl for a start because I'm familiar with it. I don't have a suitable computer with a parallel port, but I'll be looking into using LinuxCNC on a Raspberyy Pi. Plenty to research. I'll make sure I leave myself plenty of room in the control cabinet.
I made the table twice as thick and modified everything accordingly.
Will look into other motors though. Still lots to do, but wanted to give an update. I'll be sure to get some thin epoxy to reinforce where the rails mount to. Depending on minimum, quantities, I may still have some left for the rest of the frame.
Thanks for all your help,
Juan
-
Re: Small Concrete CNC Mill
Hi Juan - Your machine is very small and will not require high G's or high velocities. High G's require more powerful motors and thats $$$ look at N17 motors as they may do the job. N23 can be quite powerful and they do run quite big machines. In the commercial world you will design the machine then do a value analysis. You will determine the total build budget then apportion the cost of bits vs their functional value. For your machine you maybe over capitalizing on the motors within the project so consider N17's... so do the math to make a performance vs $$$ decision vs just picking some motors.... 0.2g is quite respectable and will change the torque requirement dramatically...Peter
4000mm/min at 400rpm is a 10mm pitch screw. You probably want a 1605mm pitch or 1204mm pitch for a good force so you won't get to 4000mm/min. Look up the torque curve of the motor.....
say you use a 1605 screw at 500rpm thats only 2500mm/min maybe 750rpm gets 3750mm/min so maybe there...
-
1 Attachment(s)
Re: Small Concrete CNC Mill
I see. If anything, I want to undercapitalize :). I did look at the torque curve of the motors, and I have done so in further analysis.
I've redone the calculations for an acceleration of 0.2g, 1605 ballscrews, maximum cutting forces of say 50 N (is that reasonable?), ignoring air drag :) and using NEMA 17 motors such as these:
https://www.omc-stepperonline.com/nema-17-stepper-motor/3pcs-of-nema-17-bipolar-59ncm-84oz-in-2a-42x48mm-4-wires-w-1m-cable-and-connector.html?mfp=145-holding-torque-ncm%5B46%2C59%2C60%2C65%2C79%5D
They have 59 Ncm holding torque and can hold about 23 Ncm up to 750 rpm. That is at 24 V, half step. They also have others with 65 Ncm but their torque curve does not go past 525 rpm.
Also 0.9ยบ motors have less torque at full step than the proposed 1.8 ones at half step.
According to my numbers they would do nicely, plus I can use something like DRV8825 stepper drivers, which makes everything cheaper.
For a power supply, according to here https://www.cnczone.com/forums/stepp...ted%20voltage.
Quote:
AMPS = SUM(amps required for all motors) X 0.66
Voltage = Minimum, 10X motor's rated voltage, Maximum, 20X motor's rated voltage
Motor rated voltage is 2.8 V, 10*2.8V = 28 V, 24 V would be fine and it is what is used in the torque curve.
Motors are rated at 2A, say I oversize the power supply in case I even want to drive 5:
5 motors * 2 A * 0.66 = 6.6 A, so the closest available 24 V power supply would be 8.2A
Juan
-
Re: Small Concrete CNC Mill
Regarding rails I'm thinking of these https://www.aliexpress.com/item/1005...c00po56al&mp=1 Their 20 mm counterparts from the same store have good reviews and I can't really afford much more.
Juan
-
Re: Small Concrete CNC Mill
Good Morning Juan - When I have a moment I'll check the maths. Well done to do the maths. I suggest you try to use a system with higher then 24V, 36V (or higher) will do better. 50N or 5kgf for cutting, will cut foam but not timber and not aluminium. Need 15kgf plus to do hard timbers. Your 5mm pitch screw with a 59Ncm (0.59Nm) motor will produce ~62kgf at stall so that's heaps....
F(kgf) = 0.59x2xPIx0.9/9.81/0.005 = 62kgf so maybe 30kgf at 500rpm (500rpm is 500x5=2500mm/min and you probably won't heavy cut at more then 1500mm/min, light finishing cuts will be as fast as machine can go) My belt router with the makita I ran at 23m/min finishing cuts.
As long as your close with the power supply it will be fine on this machine. Switchmode power supplies will only put out what they can and they are hard to blow up. If you look at unregulated power supplies you may find something as well. So break down your calcs to two conditions a) cutting which is close to constant velocity so no acceleration and b) rapids which will need to include accels. But on a small machine to move across a small table does not take much time. So you can check its accel transit and decel to see how long it will take... this is usually called a trapezoidal velocity profile or maybe triangular. Since your making many decisions based on cost downs at least you have the correct basis for making these decisions. From my experience cars and rails are important so maybe that's not the spot to cost down... or if they are really cheap then they can be upgraded into the future... As you say its a learning experience so $$$ matter. maybe since its a timber machine look at supported round rails, they are half the price of square and I'd only use square on the Z. They have a broad base more suitable for screwing to timber... but they are wider so machine gets bigger.... trade offs
also look at 12mm screws much less inertia...
Keep at it Peter
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan - Looked at your math and I haven't checked everything but the logic and most maths are correct. Well done. Now you have the drift you can go forth and design machines. This process picks a design point. It does not fully describe the machine motion. It picks a point up the acceleration curve and checks the motor has enough torque to achieve those conditions. Jerk or jounce has to be considered as well and this smooths (damps or inhibits) the motion out. It takes energy to start and stop things and change their direction so although this maths says you will get to full speed in 0.03 seconds in practice this is unlikely. If your deeply into maths or have access to dynamic FE you can calculate the actual accel time closer... But in terms of this machine design level its adequate for sizing the motors....
https://en.wikipedia.org/wiki/Jerk_(physics)
well done keep at it... for completeness check buckling as well.. Peter
be careful with spring balancing. Springs are not constant force devices. If you use a very long soft spring they can be close to constant but generally not. The idea of balancing the Z axis is to catch it if the power goes down. There is generally enough force in the system to raise and lower the unit on a small machine like this. Ballscrews can be "back driven" ie they do not have enough friction to self stall. I suggest instead of using a spring to consider a motor brake. Its easier to implement and not dear and all spring systems chew up usable Z travel. Unless you use balancing weights. Motor brakes are easy to do and may save a big crash one day!!
-
Re: Small Concrete CNC Mill
Hi Juan - Designing a balancing system for Z is actually quite an exercise in itself. Highly recommend using a brake. Simpler, no lose of Z travel etc etc. My small router with a plastic nut and 8mm trapezoidal thread will "fall down" under its own weight. Didn't do this when new but over time its worn and now I have to be careful when I turn it off. If I have a small tool in it it will break when it hits something!! Ballscrews will do this at day dot... since then all my machine designs have brakes on z. My first router I used a gas strut but that was hard finding one that was just right. They were not on spec and I had to go to the shop with scales and measure each one until I found the correct one in the shelf... Then it chewed up 45mm of Z travel. By the time you figure brackets, devices etc just get the brake! Peter
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan - the driver you are considering can go to 45V I'd go there. Peter
-
1 Attachment(s)
Re: Small Concrete CNC Mill
You have already helped me more than many university teachers. Thank you for your time, I can only hope to give this back in the future.
I've revised the calculations, separating cutting and rapids as suggested. I've included acceleration while cutting because the tool has to change direction while cutting, say, a pocket, and that means accelerating.
I've considered max cutting forces of about 200 N. Those would also influence friction in cars in addition to weight, but its almost negligible. I've checked buckling as well, not a problem for 16 mm or 12 mm screws.
With a cutting feed of say 2000 mm/min and rapid feed of 3750 mm/min, those NEMA 17 motors are fine. They will probably be able to go faster/sustain higher cutting forces, we'll see once the machine is built.
I knew about Jerk, more from T-shirts "Don't be a d^3x/dt^3" than from any physics clases, where they didn't mention it. Didn't know about snap, crackle and pop (not that they are of any use here). I suppose the CNC controller takes care of the detailed motion profile, and calculating travel time is just to get an idea, so an estimate is fine. Plus, as I said I have a plotter whose velocity and acceleration I can tune to get a feel for it, or even measure the actual time. On another note, I've got that plotter to "print" toolpaths out of Fusion 360's CAM. Not of much use, but interesting to see.
For the power supply I'll go then with a 36 V switching one. Cheaper and easier than an unregulated one. It can also be adjusted up or down a bit, say to 40 V, better than adjusting a 48 V one down to 45 just to be safe.
On the Z axis, having a relay put a resistor across the motor windings when there's no power should be easy enough and should hold the axis in place without mechanical brakes, springs or else.
The motors can move around a 16 mm ballscrew easily, so no problem. Originally I wanted to go with those supported round rails, will have a look, but I think those "cheap" linear rails would be fine, and if I need to upgrade them I'm sure they'll find a new home in another project.
I'll design a desk/table/somewhere to mount the machine and computer to as well.
Juan
-
Re: Small Concrete CNC Mill
Hi Juan - Yes there is a centripetal accel around a corner but the motion controller whilst cutting is maintaining a constant velocity. The centipedal accel is not reflected to the motor but to the tool and frame. To understand this you will need to do a free body diagram of the tool whilst cutting. But due to various things this is usually neglected at hobby level. Very fast commercial machines consider jerk. The toolpath is calculated forward, then backward then forward in an iterative manner to minimise jerk. In a valve train in your car if it uses camshafts jerk is very important a this is what makes the cam follower bounce. Then digging a little deeper toolpaths are actually linear, they are thousands of very small lines. So an arc is made up of say 0.001mm lines. Each one has a jerk at the start and its end, if the tool could perfectly replicate the toolpath. But it can't as your mechanical parts have play and elasticity. This damps and smudges the toolpath into a apparent continuous cut vs the contiguous toolpath... keep at it, Peter
going back to a statement - although I said your timber machine will cut aluminium it won't cut like a mill. You will have to be patient and figure out various things to do this. And on the arm 2 pieces of ply thick at the bearings with a good size metal washer or plate. I siggest a single think al plate.... for under the bolts
-
Re: Small Concrete CNC Mill
-
Re: Small Concrete CNC Mill
I see. Well, at any rate those motors are good enough for this application, even better if they don't have to withstand acceleration forces while cutting, I'll probably order them soon.Will read more on jerk, seems an interesting subject.
I know it won't cut like a mill. It is not a mill, more like a router in a mill form-factor. I'll go through a lot of wood before attempting any aluminium that's for sure.
I'll make the suggested changes. Thanks. Will get the plywood soon, and begin cutting out the pieces.
Juan
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Ordered motors, power supply and other bits and pieces. Ordered the rails too. Kind of blew the budget, but par for the course, things cost what they cost. Here are all the wood pieces cut out with a jigsaw. I'm using the router which will become the spindle to finish the pieces up to the marked lines. Edges where the rails will mount to will be trimmed in place when the frame is partly assembled.
Juan
https://www.cnczone.com/forums/attac...d=458160&stc=1
-
Re: Small Concrete CNC Mill
Hi Juan - very exciting, look fwd to the build. What class or type of plywood did you get? Peter
-
Re: Small Concrete CNC Mill
The plywood I got is not that great, what they had at the local home improvement store in the appropiate size. It's poplar, easy to work with but kind of soft. Now that I look at the spec sheet it is 4-5 GPa... I'll make sure I coat important surfaces with epoxy and use aluminum plate under the bolts. On the other hand it wasn't that expensive and the leftover board will become the table the machine will mount on.
Juan
-
1 Attachment(s)
Re: Small Concrete CNC Mill
All the wood pieces have been trimmed to size. The router has been great, and now I have a bag of sawdust. A couple mishaps, but nothing major and it also builds "character" :) as they say. I'll glue the "subasemblies" together, but maybe the base and column will remain separate until the rails arrive. Will be working on the electronics and the spindle mount.
Attached is a mockup. It may be a small machine, but not too much by my standards.
Juan
https://www.cnczone.com/forums/attac...d=458230&stc=1
-
Re: Small Concrete CNC Mill
Hi Juan - Looking great. I see you make model airplanes? Thats the use for the CNC? Keep the sawdust, good filler to make the next concrete machine lighter.. :) Peter
-
Re: Small Concrete CNC Mill
Quote:
Originally Posted by
peteeng
Hi Juan - Looking great. I see you make model airplanes? Thats the use for the CNC? Keep the sawdust, good filler to make the next concrete machine lighter.. :) Peter
I have that model plane halfway assembled. I turned to model submarines, less catastrophic when things go wrong (somewhat) :). This is some of my work:https://youtu.be/4K4q5dHAo6c
Will be using it for that and other projects that come up. Maybe making some PCBs as well.
Juan
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Hi Juan - You maybe interested in this. Your modelling looks to be very good. Keep Making Peter
-
Re: Small Concrete CNC Mill
Thanks, will have a read.
Juan
-
1 Attachment(s)
Re: Small Concrete CNC Mill
Interesting article. I'm sure I'll learn more on the subject on my automation classes.
Here is a simple workbench that will hold the machine and associated electronics/computer.
Juan
https://www.cnczone.com/forums/attac...d=458464&stc=1