[yoshimitsuspeed]

Hey guys. New here so hi :wave:
In the last couple days I have gotten a huge hair up my * to try and build a mini CNC.
This will actually mostly be used for aluminum and I have some hopes to do some steel but that's not required. It would just be an added bonus.
I have quite a lot of experience in the mechanical design and fabrication side of this from my last job.
I was the primary fabricator for a number of machines that we made for production in the composites industry.

Here is a picture of one of the machines. This one can lay up a unidirectional thermoplastic composite blank up to 2 meters by 2 meters.



I also have several years experience in manual machining and just a little experience with CNCs.

Anyway I got laid off and so I have started out on my own doing metal fab, welding and custom fabrication.
I would love to have a full machine shop but space for me is even more limited than money right now. Once I started thinking about it I decided a little 600x400mm CNC would be able to do a large majority of what I wanted to do. Plus it would be a great first step at getting into this on my own. Like I said I am very familiar with the mechanical side and better versed than many when it comes to the electrical but when it comes to software, programming and tying it all together well I'm still a little :drowning:

Even though I want to do mostly aluminum I decided to post here since this is where most of the DIY and most similar projects seem to be.

I wanted something very rigid and also very compact. For the sake of staying compact I decided to go with more of a gantry style than moving tables. The downside of Gantries though is you have a lot of leverage being applied to certain areas. The legs that go down to the linear ways seemed like a particular weak spot to me and after some time thinking I decided to get rid of them all together.

This design would either use 80 x 160 80/20 or 2" x 6" steel tubing.
I am leaning toward the steel tubing since it would be a little cheaper for me and should be a little more rigid. I'm thinking probably 3/16 wall.

The base would naturally have considerably more reinforcement than the drawing. Possibly 2x2 square tubing welded to the bottom then faced or something similar.
The 2x6 would sit on top of the base with rails on top of it that the gantry would sit on.



I'm not set on one idea for driving this axis. In the model the lower shafts are concepted ballscrews. I wasn't sure if belt drive would be best or if I could put a motor on each one and if that would be best. Or if I could do something like a gear drive on each.
I am open to suggestions on that. I like the idea of two ballscrews because between those and the linear rails it should be quite rigid and resistant to torsional forces in the x,y directions.
One other thought would be to run a second gantry rail across so the z axis was supported on both sides. I don't know if this would be necessary but it would make for one burly little unit and it wouldn't really get in the way much.

So I guess that's pretty much it for the mechanical. I'd love any input I can get.


I am also curious what sized and type motors people would recommend for a setup like this, what kind of controllers should be used and how to tie the whole package together properly. I'm not looking for crazy feed rates or anything.
I am almost completely Linux based and would prefer to run it off linux if possible. From what I have read it seems like there are some decent options out there. I just don't know if I need to pick specific controllers/hardware with that in mind or if it will work with anything.
Of course I want Ferrari performance but more importantly I'm on a pack of ramen budget.

[wingarcher]

Your belt drive as drawn will have difficulties getting sufficient tension on the belt to get the backlash out of it. Seems like slapping a pair of motors on (one on each ballscrew) is pretty simple.

No need for a pair of gantries.

N

[louieatienza]

You probably don't want to run a timing belt that way since there is not enough tooth engagement with the belt and pulley. Then again, if you want stiffness, you might want to couple the ballscrews directly to the motors for better stiffness. This would also free the ends of the machine up since you don't have a belt running across.

As to the ballscrew locations, there's another thread where we discussed how having the ballscrew coplanar with the linear bearings can increase stiffness.

There doesn't seem to be much detail on the mechanicals of the Z axis. But, if you look at some builds here you'll see that the bearings on the gantry ends are spread out farther, to help support the cantilevered spindle. This does come at a cost of machine travel, but since the spindle is limited in travel by design, it might be a good idea to widen the gantry ends.

Speaking od useability, You may find it a little bit of a pain to have to reach into your machine to load, set up, and remove parts.

[yoshimitsuspeed]

Thanks for the replies.
The belt setup was just a quick implementation of general theory. If I went that direction I would add some idler pulleys to increase contact on the motor. Honestly I would rather get away from the belts or other drive setups and run a motor on each if that wasn't too much trouble.
I don't have a lot of detail on the Z axis at the moment but there is a lot more detail that would have to be added in general before the final iteration. The other thing is I am using freecad which is still pretty new in development and I am still learning the program so my abilities are still more limited than they would be with a better CAD program.

Again for the same reason I put the ballscrews where they were quick and easy to model but in my mind I did have them much closer to the plane of the linear bearings.
Any way you cut it this seems like it should be more rigid than having long legs on the gantry. It may make access a little harder but I don't see it being a significant issue.
I do like the fact that it would cause the chips to be more contained and easier to contain.
While modeling it I even had the idea that with this setup I could just flood the bottom with coolant or easily fit a coolant recirc system. The pros for me seem to outweigh the cons.

I appreciate the input and hope to get more info and opinions.

[dodger889]

Here is a thought for you you would like to be able to do steel too.

How thick are you thinking for doing?

If you are doing plate type of machining. You might want to use a plasma cutter instead of milling the steel. Which would require you to have two mounting system for your z axis. And a different bed that you could change out.

Now remember this is just a thought on my part take it for what it is worth. (group)

[nickcornilsen]

It sounds like you're going to want this machine to make money while you're laid off (Or, better yet... make money working for yourself and never have a boss again!)

Given that, put a dollar amount on your time. I value my personal time at 32 bucks an hour, and my professional time at ~95 an hour, for reference.

Now, estimate how much time you'll put into designing this machine. Then, consider the likely case that what you build will be a compromise for a number of reasons.

Then, price used equipment that meets your needs and satisfies as many of your wants as possible.

I would bet dollars to dimes that the used equipment will be cheaper... buy it, and bill as many of your hours to customer projects as possible!

If you want to do this a a hobby... go for it.

As for your design - I've found that in gantry style equipment, you get far more bang for your buck for a given vertical work envelope by increasing the space between linear bearings (on each side) than by reducing the height of the gantry. But I'm no fan of gantry style machines for a number of reasons.

[OCNC]

So I guess that's pretty much it for the mechanical. I'd love any input I can get.

This looks a lot like Bill Jesson's modified Momus.

Chris

[yoshimitsuspeed]

Here is a thought for you you would like to be able to do steel too.

How thick are you thinking for doing?

If you are doing plate type of machining. You might want to use a plasma cutter instead of milling the steel. Which would require you to have two mounting system for your z axis. And a different bed that you could change out.

Now remember this is just a thought on my part take it for what it is worth. (group)

Yes I had thought of that. Most of what I would do with steel would be able to be done with a plasma. It seems like it would be pretty easy to have an insert bed that kept the main surfaces clean.

[dmalicky]

Welcome to the forum. Good advice above. Sounds like you have good experience for building a machine.

If your primary aim is to cut aluminum, you'll need a stiff machine. This thread:
http://www.cnczone.com/forums/diy-cn...tml#post948730
says the desired range is 20k to 50k lb/in. I would guess most DIY machines are similar to the ones in the thread -- 2k or less.

To get it stiff, a fixed gantry / moving table design is the way to go. It does take more space, but the gantry can easily be made both rock solid and heavy. As you know, mass is good for machining, but a heavy moving gantry will either be very slow or require expensive motors/drives.

If you're really tight for space and don't mind very slow, then I would first look at some successful commercial machines for their approach:
CNC Router - MultiCam - Products - 3000 Series
Haas GR-510 | Haas Automation, Inc. | CNC Machine Tools
- The gantry uprights (legs) won't be a major source of flex if they're made of a stout box section, and aren't too long. I did some FEA models that show the same thing.
- The main rails should be as close to the cutting plane as possible. Those commercial machines put them just under the table for better user access. But just above the table--in the same horizontal plane as the cutter tip--is actually best for stiffness. At that height, cutter loads do not cause any tipping of the gantry. Higher than that, the gantry tips the opposite way (compared to below the cutter tip).
- Quote:"Honestly I would rather get away from the belts or other drive setups and run a motor on each if that wasn't too much trouble." Yes, dual motor drives on the main rails is much stiffer than a belt or even chain, and isn't difficult. Ideally, the leadscrews are on that same plane as the cutter and main rails, which is easily done with dual screws.

But again, a fixed gantry is much better for metal cutting.

[yoshimitsuspeed]

It sounds like you're going to want this machine to make money while you're laid off (Or, better yet... make money working for yourself and never have a boss again!)

Given that, put a dollar amount on your time. I value my personal time at 32 bucks an hour, and my professional time at ~95 an hour, for reference.

Now, estimate how much time you'll put into designing this machine. Then, consider the likely case that what you build will be a compromise for a number of reasons.

Then, price used equipment that meets your needs and satisfies as many of your wants as possible.

I would bet dollars to dimes that the used equipment will be cheaper... buy it, and bill as many of your hours to customer projects as possible!

If you want to do this a a hobby... go for it.

When I worked full time i fully would have agreed with you on the value of my time. Currently I have much more time than money so the scales balance much differently now.
On top of that I have skills and knowledge that are slowly fading away. I would rather keep those skills sharp and progressing instead. This has a good bit of inherent value to me.
I am starting my own business doing custom fabrication so this would be a nice little asset to have. My shop however is a 26' Uhaul so space is increadibly valuable. As is ease of mobility of everything inside it.

As for your design - I've found that in gantry style equipment, you get far more bang for your buck for a given vertical work envelope by increasing the space between linear bearings (on each side) than by reducing the height of the gantry. But I'm no fan of gantry style machines for a number of reasons.

I assume you mean having two bearings per side and increasing the space between them right?
This was what led me to my though on removing the legs. It's not the rigidity of the legs themselves but the fact that is seemed to me any play in the system would be multiplied much more the longer the legs were. If you had one thou play in each bearing allowing the gantry to move back and forth on the vertical axis that would translate to much more movement at the top of the gantry than if the bearings were on the same plan as the gantry. Moving the bearings further apart would also greatly reduce this motion but more than 2" from the centerline of the cutter and I have to increase my footprint or reduce my cutting area.
I agree that gantries have their downsides but I have seen them work perfectly well and know it could for my needs.

[yoshimitsuspeed]

Let's talk a little more about motors and controls because this is where I am the most lost.
What sized motors do I need? From what I understand steppers are the cheaper alternative so I'm guessing that's what I'd go with. I'm not looking for crazy speed but need to be able to maintain good feeds and speeds for various materials and cutters.


Since I prefer Linux would linuxcnc be the way to go or is there a better route?
When choosing electronics do I have to consider this or will any program be able to run any setup?

What are suggestions for a spindle? It's likely I would start out with a router but as some point I would like a variable speed spindle.

[yoshimitsuspeed]

Would something like this be an acceptable starting point?
4 Axis NEMA 23 Stepper Motor 290oz in 4 Axis Board CNC Kit | eBay

And how about something like this for the spindle?
I have no idea what I would need as far as size or power. I don't see myself going over 1/4" mills though.
300W CNC Spindle Motor Kits Power Supply Speed Controller Support Holder Motor | eBay

[dodger889]

Well this could get interesting. Since you have a little knowledge of the electronics you have two choices. Buy a complete setup for one source or try to piece meal one together and save a lot or cost you of money . Now which are you looking at going? Trail and error or the sure thing? The trail and error could save you a lot. But could double your cost. Also you will need to know which way are you going to drive the mechanics of the machine. Ball screws, ring and pinon, acme screws, or even threaded rod. This all plays into how or why you would chose the drive motors and drivers.

[dodger889]

Would something like this be an acceptable starting point?
4 Axis NEMA 23 Stepper Motor 290oz in 4 Axis Board CNC Kit | eBay

And how about something like this for the spindle?
I have no idea what I would need as far as size or power. I don't see myself going over 1/4" mills though.
300W CNC Spindle Motor Kits Power Supply Speed Controller Support Holder Motor | eBay

Both are on the very light side of what you are wanting to work on. The spindle I would not waste my time on. For what you said you wanting to do the least for the spindle would be in the 2kw ( 1.5hp) or more power. The motors could work at a more cost factor for gear reduction system ( pulleys and timing belt) this could get you at least double the power.

[yoshimitsuspeed]

Ball and screw seemed the most practical to me but I am open to other suggestions.

What size motor should I be looking at for direct drive of ballscrews on the X,Y and z?
Don't the ballscrews gear the motors down pretty substantially?

[dodger889]

Ball and screw seemed the most practical to me but I am open to other suggestions.

What size motor should I be looking at for direct drive of ballscrews on the X,Y and z?
Don't the ballscrews gear the motors down pretty substantially?

That would depend on what threads per inch you are going for. More TPI the better the gearing ratio but at a cost of speed. The opposite is true for less TPI the less gearing effect but with greater speed . Then you need to figure out what is a happy median you are looking for.
I have some ball screws that regret now after getting them. I wished I had not. For one they are way too fast for what I need. One set is 1.5 inches per turn the other is almost two inches per turn. So I will put them up on fleebay and hope to get most of my money back. And they are C3 or C4 screws....

[louieatienza]

Ball and screw seemed the most practical to me but I am open to other suggestions.

What size motor should I be looking at for direct drive of ballscrews on the X,Y and z?
Don't the ballscrews gear the motors down pretty substantially?

extremely high efficiency.

Depends on what ballscrews you get. Theoretically, you'd need less power from the motor to turn ballscrews as opposed to leadscrews because of their extremely high efficiency. This may or may not cause problems for the Z axis, as it may drop when the power is off.

If you're using steppers like most, you'll have to balance speed and resolution, since you'll lose torque as speed increases, and you'll need the feedrates to machine aluminum. You may not be able to run lower speeds unless you also use a milling spindle that has sufficient torque at lower RPMs.

I would check out the GeckoDrive website, There are some really imformative documents and formulae there to help get you started. But I'd suggest you build you machine first, then once you know the weight of yoru components, you'll be better able to figure what pitch screw and what size motor is needed to move them at the speed you need to for machining. If you plan on using a milling head and slower feedrates, it may not be worth it to get ballscrews.

As to LinuxCNC, it is free and pretty robust. Theere have been compatibility issues in the past, but I don't know of they have been resolved. I believe that at my last look, there is not USB support for LinuxCNC, and that may affect your purchase. I would say there are far more Mach3 users so help would be easier to find. Sherline actually ships LinuxCNC with their CNC machines.

[yoshimitsuspeed]

This seems so weird.
If I could just find max RPM or even better torque/RPM graphs for the motors and thread pitch on the ballscrews I could start figuring out what kind of gearing I want. Very few of the motors seem to have that info and I can't find any ballscrews that have it. Maybe I'm just not reading the charts right. Many of the ballscrews have a specs page similar to the one linked to on this page. Block and Rail in USA. They have a ton of specs but none of them look to me to point to thread pitch. Am I just missing it?
Block and Rail in USA

[Trotline]

Pretty sure the second number in the spec. grid for the ballscrews is their lead, in mm. Those come in 5mm lead, (5mm per revolution), and 10mm. The shorthand for that is in the assemblies, where a 16mm ballscrew with a 5mm lead is described as a 1605 set. The C7 at the end of it describes their tolerance. There's a dynamite thread about all this.

http://www.cnczone.com/forums/linear...ew_basics.html

Cheers!

Luke

[louieatienza]

This seems so weird.
If I could just find max RPM or even better torque/RPM graphs for the motors and thread pitch on the ballscrews I could start figuring out what kind of gearing I want. Very few of the motors seem to have that info and I can't find any ballscrews that have it. Maybe I'm just not reading the charts right. Many of the ballscrews have a specs page similar to the one linked to on this page. Block and Rail in USA. They have a ton of specs but none of them look to me to point to thread pitch. Am I just missing it?
Block and Rail in USA

It's tough to know the max RPM for a stepper since there are many ways it can be configured, many different drives they can be hooked up to with different current and woltage that can be applied to them. Without any load, a stepper can spin over 3000rpm with the right drive and controller, but will have very little torque. From the machines I've built and the ones I've seen moving, you probably have a range of about 0-800rpm before the torque falls off too much to be useful. Adding more voltage will allow the stepper to go faster before the torque falls, though it will ge hotter.

The screw diameter is the first column, the lead is the second column. Note that the pitch and lead of a screw can be different. The lead refers to the distance the nut travels per one revolution. The pitch refers to the distance between the tops of the threads.

I would check the reference papers at Nook's, THK's, Thomson's, and Roton's websites for more detail on leadscrews and ballscrews. THK and Thomson have extensive documentation for their screws. Nook has a screw speed calculator as well as other tools...