Another n00b wants to build a CNC machine!
Hey everyone,
Jason here, from Cedar Rapids, IA. A couple of years ago, I bought and assembled my first 3D printer out of a kit. I had a blast building and using it, andI learned a lot, too. Now a few weeks ago, I built my first CNC machine (made out of 3D printed parts, 3/4 inch EMT conduit, and an Arduino controller, anda handheld Dremel tool as the spindle. This project was even more fun and it worked surprisingly well, even taking very thin cuts out of aluminum. I think it's safe to say that I've been bit by the bug! So here I am, deciding that I want to make an even BETTER one. I have an old heavy steel desk, about 1500mm by 860mm, and my idea is to permanently build the machine onto that. I plan to use this machine for milling small aluminum parts, and larger wood projects. I don't need insane levels of accuracy, but getting to half a thousandth over a foot distance would be pretty nice. I started with one of those Chinesium 2.2kw water cooled 80mm spindles, with a VFD. I tested it out, and it seems to work fine. My next task is choosing linear parts, motors, and such, That is what brings me here. I've been looking at parts online, and I have found a few things on Amazon that LOOK like a good choice to me.
Here are the steppers I'm looking at..
LINK
Three nema 34 motors, with individual power supplies and DM860a controllers. This looks good to me. I can handle the price, and I THINK they're bit enough. Does anyone think these are too small?
The next thing I'm looking at is rails. I have found plenty of different options to tinker with, but I know little about what is best suited to the task.
Option 1
Here is everything bundled into a kit, which is tempting. The screws are C7, which (judging from what I've found on this site) isn't really that awesome, but It's probably more accurate than I am.. But my gut is telling me not to trust this style of rail, so I've also been looking at rails of a different style..
Option 2
These "feel" better to me, but I really can't explain what makes them better. I think that, by being lower profile, they won't be as likely to "give" when I'm putting who-knows-how-much-force on them. Obviously, I would need to buy individual sets, instead of having a kit.
So if you're still with me, you have probably figured out that I'm setting out for a project that is slightly above my pay-grade, but I figure knowing that is half the battle. I'm asking for thoughts, opinions, constructive criticism, suggestions on what might be a better choice of components. I fully plan to keep this thread updated as I proceed with the build, so you guys can have pictures and such. Maybe it will be a good writeup for someone who comes along.
Jason
Re: Another n00b wants to build a CNC machine!
I decided to go and buy the stepper motor kit from Amazon. Sold by Long's Motor, this consists of 4 nema 34 motors (1600 oz/in), 4 DM860a drivers, and 4 power supplies. I bought 4 so I would have a spare of each. cost was $640, free shipping (not prime). I'm assured that these parts are made of 90% pure Chinesium, and are guaranteed to work correctly until connected to a power source.
Jason
Re: Another n00b wants to build a CNC machine!
I was going to say Do NOT buy those motors. Cancel them if you can. Unfortunately I'm getting on a plane right now but I'll reply again later tonight.
Re: Another n00b wants to build a CNC machine!
Quote:
Three nema 34 motors, with individual power supplies and DM860a controllers. This looks good to me. I can handle the price, and I THINK they're bit enough. Does anyone think these are too small?
With steppers, bigger is rarely better.
A steppers rated torque is the holding torque, when the motor is not spinning. As rpm increases, torque decreases. generally, the larger the motor, the faster the torque decreases. Because of this, it's very common for a motor in the 300oz range to have the same torque as those 1600oz motors, when you get over 600-800rpm.
When you look at the ballscrews in the link you posted, they are 5mm pitch screws. So they need to spin 5 times to move 1 inch. That's 500rpm to move at 100ipm, which is quite slow for a machine that size. 5mm pitch screws are generally a poor choice for a router, and when you combine them with 1600oz motors, you have a worst case scenario. You can find build here where members went from 1600 oz motors to 400 oz motors and had huge increases in performance.
For a machine that size, low inductance motors in the 400oz range will provide the best performance, when combined with high quality drives running at about 60V. Low inductance is key, as that dictates how fast the torque falls off when rpm's increase.
For the screws, you want them to have a 10mm pitch. This doubles (or more) your speed by spinning the motor slower, where it has more torque.
Quote:
Option 2
These "feel" better to me, but I really can't explain what makes them better. I think that, by being lower profile, they won't be as likely to "give" when I'm putting who-knows-how-much-force on them.
Profile linear rails are much more rigid than the round rails, as they have a far greater load carrying capacity. One issue with the ones in your link, is that 15mm rails are quite small in size, and can be harder to work with, due to the small screw sizes. 20mm rails are actually a lot larger, and much easier to work with, but a bit more money.
I'm not a fan of the round rails at all, and would not recommend them. Several members here that have used them have found them to be the weakest link in there machines.
A lot of those chinese package "deals" seem like great bargains, but the performance they deliver is on par (or worse) than their low prices.
Re: Another n00b wants to build a CNC machine!
Well you've bought them now so you might as well use them. I use a smaller DM542A driver from Longs motor and I'm very happy with them. A lot of routers use rack and pinion drive. I use this 5:1 timing belt geared drive https://www.aliexpress.com/store/pro...32bf3ef74VlzTd
and with some small NEMA 34's I am pulling 825 inches per minute on my gantry (plasma).
I think for your sized machine, you might have trouble with ballscrew whip limiting travel speed unless you use a rotating nut design.
I agree with using linear rails and while I started with 2 x 15mm rails on my gantry which I bought from eBay locally, I then used 25mm rails which I bought from aliexpress for about the same price I paid for the 15mm. I used the 25mm rails with those geared drives up and down the table (1600 mm rails give me about 1300mm of travel) . If I built another ganty I'd do it with 1 x 25mmm linear rail. A lot of people will tell you you need spring loaded pinions to manage back lash but when I asked commercial builders, the said it was not necessary if you can build an accurate machine which I did. I've never had problems with backlash.
Re: Another n00b wants to build a CNC machine!
rodw and ger21, Thanks for the replies, and good to meet you both. As it turns out, checking my email this morning shows me that the motors have already shipped, so I'm probably stuck with them. Thanks for confirming my hesitation about the rails. I can't always put my finger on exactly why I don't trust something, but what you say fits perfectly with what I was thinking. Taking from what has been suggested, I'm searching again, and here is my next guess for rails and screws, bringing the 10mm pitch into account.
1500mm linear rail for the long axis, for $270 along with This ballscrew for $203
1000mm linear rail for the short axis, for $225 along with This ballscrew for $161
400mm for the Z axis, for $150 and this ballscrew for $70
Which comes out at just over $1000, which is possible. It's more money than I was hoping to spend, but I imagine this will be a much more rigid setup than the round rails, while being easier to work with during the build. I WAS toying with the idea of using a rack and pinion method for the long axis, but I haven't yet dug much into it. Maybe I need to give that some more serious thought. It seems like my mistakes so far can be chocked up to lack of experience. My little 3d printed cnc dremel (toy) only cuts at about 20mm per second (about 47 IPS), so I've never really known what "normal" speed is. I guess I never imagined moving at much higher speed for anything other than moving from one target to the next between cuts. It makes me wonder what speeds I should actually be looking for.
Rodw, I'm not sure what you mean by rotating nut, but I THINK you mean holding the ballscrew stationary and rigging the nut to spin, pulling the carriage up and down the axis. Am I reading you right? And by ballscrew whip, you mean the screw turning so fast that it starts to bow out in the center like a jumprope?
ok, Lessons learned today:
Bigger steppers don't mean better, because while all steppers lose torque at higher RPM, bigger steppers lose MORE, and lower induction steppers lose LESS.
Round rails suck and nobody is really all that happy with them. The low profile rails are more rigid, and it's usually worth spending a little extra to go a size bigger for something that's easier to work with.
As always, I will happily listen and learn from anyone who wants to chime in!
Jason
Re: Another n00b wants to build a CNC machine!
Quote:
Originally Posted by
nr0x
Rodw, I'm not sure what you mean by rotating nut, but I THINK you mean holding the ballscrew stationary and rigging the nut to spin, pulling the carriage up and down the axis. Am I reading you right? And by ballscrew whip, you mean the screw turning so fast that it starts to bow out in the center like a jumprope?
Jason
I've got no personal experience with rotating nuts or long ballscrews but from what I've heard on longer ball screws the speed is limited by "whipping" of the shaft. You can buy rotating nut ballscrews but pretty expensive. Some people have made their own. Basically, you bolt a timing belt pulley to the flange and encapsulate the nut in a thrust bearing. Some have suggested even a cheap car wheel bearing. BUt it does imply you need some machining ability.
https://youtu.be/4unEngqU09A
Re: Another n00b wants to build a CNC machine!
Spend at least a month doing research before you spend any more money. It may save you a lot of money.
Re: Another n00b wants to build a CNC machine!
That I will certainly do! The motor set is here. I haven't had time to test them out as of yet. The first thing I noticed is that these stepper drivers seem to be capable of extremely small microsteps. At first, that seems cool and "super precise", but am I correct when I think that the motors have lower holding torque with smaller microsteps?
Always more to learn!
Re: Another n00b wants to build a CNC machine!
Quote:
Originally Posted by
nr0x
That I will certainly do! The motor set is here. I haven't had time to test them out as of yet. The first thing I noticed is that these stepper drivers seem to be capable of extremely small microsteps. At first, that seems cool and "super precise", but am I correct when I think that the motors have lower holding torque with smaller microsteps?
Always more to learn!
In my experience, torque is unaffected with microstepping on these drives. I am running 20x microstepping on mine and can get 21 metres a minute rapids. I think most stepper controllers today have some smart internals and I don't think the pulses given to the motor are always what you give the driver with modern stepper drivers.
Re: Another n00b wants to build a CNC machine!
Part I (particularly the section on guide systems) of CNC Machining Handbook, by Alan Overby, would be very helpful for you, I think, as well as the appendix: engineering process of selecting a ball screw.
Re: Another n00b wants to build a CNC machine!
Well crap! I thought I'd posted an update already! My Bad!
First, thank you jo_ky for the book recommendation. I learned new things, un-learned bad things, and confirmed some of my ideas. Thank you, Sir! Among other things, I've learned that my idea of using this old metal desk to build the machine on just isn't going to cut it. As such, I'm designing a table out of 2 inch square steel that will bolt to the floor in a way that allows the height of each corner to be adjusted for truing it up. I also found a decent machine shop about 15 minutes away, and I'm bouncing ideas off them as I design. The 20mm rails I ordered are here, and I'm glad I listened to you guys and went with these instead of the 15mm ones. These will CLEARLY be easier to work with, an are surely sturdier.
So at this point, I'm looking at design considerations and looking at half a million pictures of other tables, and I'm starting to have a picture in my head of what I want to build here. I'm picking up little things, like where to mount the rails so they don't get covered in chips and crud. I've also taken a closer look into driving the long axis with rack and pinion instead of ballscrew. I believe I am capable of fabricating the parts for the pinion, based on what I have seen other people do. I've also come to see that my original ideas weren't really going to be rigid enough, so I'm making a lot of changes in the grey matter between my ears.
But here's a "what if" question that you guys might know something about. What if I build the truss across the long axis instead of the short one? Is there a reason that's just a stupid idea?
Jason
Re: Another n00b wants to build a CNC machine!
I'm assuming you mean the gantry when you say truss?
One reason, is that the longer it is, the more it will flex. Some people do build machine with the gantry as the longer access, due to space constraints, or better access to the work, when the machine is used for a specific application.
Re: Another n00b wants to build a CNC machine!
Gantry! That's the word I was looking for! Here are my thoughts..
I think that, no matter which way I build it, I'm going to have to drive that gantry from both ends, otherwise the whole things is gonna shift and strain the crap out of the linear bearings, or jam up the whole thing. So that means I have to buy double the ballscrews or rack & pinion parts. It would be cheaper to do that on the short axis than the long one. Also, the machine is going to have the long axis against the wall, and it will be between my lathe table, and some shelves, which means I'll be standing on the long axis while I'm using it. Doing it this way would probably make it easier to use.
So my understanding now is that it's possible, but generally not a good idea because it could flex more. I wonder if I can beef it up to compensate. I have a few friends who are mechanical engineers. Maybe I can get them to help me figure out just how much force will be on this thing and how much flex I can anticipate.
J
Re: Another n00b wants to build a CNC machine!
Checking in. I needed a really straight edge to help with this project, and I came into a 5 foot long chunk of 3 inch I-beam. I found a machine shop about 20 minutes away, and hired them to mill one side flat for me. When I went to pick it up, we got to talking about the project I'm working on, and my interests, and he offered me a part time job on the spot! So besides my normal job, I've started working Saturdays working with actual proper CNC machines. of course, I'm WAY more excited about this than a respectable adult should be, but I never claimed to be respectable in the first place. I'm learning a LOT!
In the meantime, I've been working with Fusion360 on designing my table, and I think I'm starting to turn a molehill into a mountain. I have this big crazy steel framework and all kinds of goofy stuff that a good quality machine should probably have. But this is my first real build here. Maybe I need to dial it back a little bit and go with something simple. I found a machine junkyard, and I'm looking into the possibility of buying a big steel t-slotted mill bed. It's about 2 1/2 by 5 feet, about 4 inches thick, and it's probably flat enough for a hobbyist like me to get decent results out of if I use it as the base for the table. Bolt the linear rails directly to that, instead of trying to true up a welded nightmare of a homebrew table. Sure, maybe I'm not gonna get 2 tenths out of it, but I'm not making space shuttle parts here. I'm dinking around in my garage. it also seems easier to bolt the rails to the top of the table than the side. I mean easier to drill 2 sets of parallel holes on the same surface than to make them on the sides. By doing so, I'm skipping the overhang to help keep chips out of the rails. Maybe I can find some kind of baffle to shield them with instead.
Anybody care to chime in and tell me where I'm about to screw it up? Seriously, cuz I'm pretty much winging it here! :devious:
Re: Another n00b wants to build a CNC machine!
Oh, and I got a little distracted by where to put this in my workshop, which I eventually decided is just too damned small. So I ripped out the wall between it and my garage, and just took over the whole place! The machine won't be up against a wall anymore, so there's no reason to try and build the gantry on the long axis, so I can scrap that noise, too.
