[oveaune]

Been thinking about making a CNC for a while now, and finally got the time to begin designing the frame, but have some questions first.

I was thinking about building the frame in aluminium, but what thickness would you recommend ? The CNC will primarily be used for PCB, carbon and aluminum

[wizard]

There are some good and very recent threads already covering this issue. Due to the wealth of information I'd suggest searching for those threads.

Been thinking about making a CNC for a while now, and finally got the time to begin designing the frame, but have some questions first.

I was thinking about building the frame in aluminium, but what thickness would you recommend ?

That means what exactly? For us it means nothing because we don't know what you are talking about. It could be sheet metal, castings, extrusions or bar stock for all we know.

Given that I'm partial to steel myself, mainly because it is often much cheaper.

The CNC will primarily be used for PCB, carbon and aluminum

The problem here is that this could mean dramatically different machines are required. After all a PCB mill can sit on a desktop whereas a machine to do aluminum sheet goods could be very large in comparison. Given that I might suggest a PCB mill as your first machine due to a lower cost and simpler construction. The idea being to see it as a learning machine.

[oveaune]

I'll try scanning the forum for simular threads, sorry should have done this before posting...

Saw by reading my previous post that I "might" have given a bit short explanation When I said aluminium, I ment aluminium plates that I cut from and make up the frame. I always thought the other way around, that steel was more expensive, but I'll check prices again. If steel is cheaper there no point doing it in alu

Good point reguarding the PCB VS Alu machine size, but the alu machine will most likely be "benchtop" size to, not a biiig machine

[wizard]

I'll try scanning the forum for simular threads, sorry should have done this before posting...

A little while ago the thread was about three quarters of the way down the first page. There is actually a couple of threads but one has an excellent explanation by a regular here on the forums.

Saw by reading my previous post that I "might" have given a bit short explanation When I said aluminium, I ment aluminium plates that I cut from and make up the frame.

This is likely a very bad approach to machine design. Obviously I don't know what your intentions are nor your design experience but building from plate leaves a lot to be desired in my opinion. This especially the case if you have limit shop facilities.

I always thought the other way around, that steel was more expensive, but I'll check prices again. If steel is cheaper there no point doing it in alu

Actually that is very much an ongoing discussion. Steel might be cheaper in my case because I can pick up stuff at discounted prices or even free in some cases. One local supplier sells cutoff, and other pieces of steel from around a dollar to a $1.60 a pound. That isn't really bad at all. Of course building a big machine from cutoff, end pieces and the like isn't really doable. However for a small machine it can make sense.

As for your design I'd recommend a cut list that covers both a steel design and an aluminum one. Price out the materials and go from there.

There are other considerations, for example a steel can hold thread better than any aluminum in my opinion. Further it is far easier to weld if you have to have something welded up or DIY a weld. Neither material as the favorable properties of cast iron but we are talking reasonable expenses here. One thing steel does do for you is add mass faster than aluminum does. In the right place mass is very useful in most machine tool designs.

Good point reguarding the PCB VS Alu machine size, but the alu machine will most likely be "benchtop" size to, not a biiig machine

Still in this case you will have to produce a far more rigid machine than is required for PCB machining. That isn't a bad thing really.

Speaking of rigid if you are going to seriously machine aluminum of any metal for that matter you need to pay attention to machine rigidity. In one of those other threads an engineer goes into far more detail than I ever could with respect to the design of the gantry beam. In simple terms cross-section wins. So steel or aluminum you get the best results from a big square or squarish beam mad from tubing.

******************* a PSA ****************

If you expect to do a lot of PCB, fiberglass or carbon composites seriously thing about a totally enclosed machine to keep the nasty dust at bay. Your lungs and body will thank you for it. This does add expense but you don't need to get excessively fancy here either.

[dmalicky]

Good advice and questions by wizard, as usual. For machine stiffness from an engineer's point of view, here are some posts to start with:

The big picture with all components: http://www.cnczone.com/forums/diy_cn...ml#post1408740
More big picture -- Stiffness-at-the-tool is a mainly a function of these factors: machine configuration (moving table vs gantry), gantry length, Z clearance, component cross-section shapes and sizes, materials, joint quality, relative leverage arms (mostly the rail and bearing locations/heights/spacings), and bearing types. All of those have to be done well for the stiffness-at-the-tool to be high -- i.e., the weakest link will be exploited the most.

If you don't mind some math, here's a post on steel vs alum for the gantry cross-member, and why tubes are generally better than solids: http://www.cnczone.com/forums/diy_cn...ml#post1419700

This post looks at what kind of tube and how big to make it (for a 4' gantry with 10" Z): http://www.cnczone.com/forums/diy_cn...ml#post1413406
and more posts with more variations down that page.


Materials
For the lower frame, I find steel is usually best: stiff, cheap, heavy, and weldable (most joints). Plywood or MDF can be good choices for lower-cost machines without a welder. T-slot alum extrusion is easy and straight, but not all that stiff, unfortunately light, and quite expensive.

For the gantry tube (assuming it is moving):
- Small machines can use steel (or alum) tube, since the gantry is light by virtue of its size. Or the largest T-slot alum extrusions, since it's convenient, stiff enough, and inexpensive for a short gantry. For a very short gantry, solid aluminum may be ok if very thick.
- Large machines are usually better off using aluminum tube, so that the reasonable cost motors & drives can accelerate it quickly.
- For any size machine cutting wood, or a small machine cutting alum, Baltic Birch plywood is cheap, stiff for its weight, easy, and provides some damping. The cross-section needs to be BIG to makeup for the lower modulus.

For the gantry car: I like aluminum tooling plate for flatness, thickness, ease of machining, and acceleration. Steel could work if flat enough, but acceleration may suffer. A 'channel' cross-section is often needed for stiffness and packaging. It's difficult to make BB ply work well here: the forces are just too high and concentrated, and the space small.

For the Z car: ditto and a channel-section closely tied into the spindle is usually required for the same reasons.


Once you get the gantry, cars, and frame stiff -- unless using profile rail/bearings -- the next major source of flex will likely come from the linear bearings. Here's a starting point for that:
http://www.cnczone.com/forums/diy_cn...ml#post1426678

[scott0999]

I'm building a machine based on 800/500/300mm SBR rails

at first I was going to do a moving gantry, but decided against it. building a fixed gantry entirely out of steel now. Some of the Z axis parts may be aluminum, still undecided on that. at this point it seems I might as well go all steel

Steel is a good choice for a fixed gantry. It's cheap, you can build it strong & heavy without worrying about too much weight. My machine will have a fairly narrow Y-axis, so using 1 X-axis motor. If I built it with a moving steel gantry, that would weight a bit for a single X-axis Nema 23 to handle. (Not one that would match the other motors well anyway)

my recent thread: http://www.cnczone.com/forums/diy_cn...thickness.html

[oveaune]

Been reading lots of build threads and it making moving gantry makes it very difficult to get the frame stable enough to mill aluminium and steel.

So I think I've decided to go with fixed gantry. Having the X axis on moving table, and Y and Z on the gantry, or also moving the Y to the table, but having only X on the table will give a smaller footprint of the machine. If building on steel the gantry should be stable enough to house the Y axis I would think ?

I cannot weld, just cant never goes well So thought perhaps to see if there is any way to put the frame together using bolts, could be difficult with steel profiles ( bolting together I mean ), and using steel plates would probably result in a weaker frame...

Perhaps making the baseframe in steel plates and using steel profiles to build up the Z axis. Steel plates can perhaps be used for the Y-axis as well.... Guess I can simulate the deflection in solidworks..

[wizard]

Been reading lots of build threads and it making moving gantry makes it very difficult to get the frame stable enough to mill aluminium and steel.

So I think I've decided to go with fixed gantry. Having the X axis on moving table, and Y and Z on the gantry, or also moving the Y to the table, but having only X on the table will give a smaller footprint of the machine. If building on steel the gantry should be stable enough to house the Y axis I would think ?

It is all about getting the right cross sections for the part of the machine in question. Simply building out of steel does nothing for you if you don't have the right sections in the right place.

I cannot weld, just cant never goes well So thought perhaps to see if there is any way to put the frame together using bolts, could be difficult with steel profiles ( bolting together I mean ), and using steel plates would probably result in a weaker frame...

It might not be as bad as you think. Welding would certainly be better in some cases but welded parts often need post processing machining. For a bolted together machine you might get by with a good bandsaw and a 4" grinder.

In general bolted together frame members are possible and does have the benefit of no welding induced stress. In some ways it might be seen as more work. A bolted together structure would greatly benefit from being doweled pinned after the machine is aligned to help keep it in alignment. The big problem here is that dowel pins really need some meat to latch into.

A more rational approach might be to use the talents of a professional welder and only have critical parts welded up. You will likely need a stress relief and some final machining though. In the end you add a few hundred dollars to the project.

Perhaps making the baseframe in steel plates and using steel profiles to build up the Z axis. Steel plates can perhaps be used for the Y-axis as well.... Guess I can simulate the deflection in solidworks..

Plates are a problem. They need to be rather thick to replace steel tubing and thus become costly. Further the structure can become heavy though that isn't a bad thing for a machine but can be a bad thing for the builder.

[oveaune]

It is all about getting the right cross sections for the part of the machine in question. Simply building out of steel does nothing for you if you don't have the right sections in the right place.

Could you please elaborate on this ? Just so I'm getting it correctly ( English not my primary language ).

Plates are a problem. They need to be rather thick to replace steel tubing and thus become costly. Further the structure can become heavy though that isn't a bad thing for a machine but can be a bad thing for the builder.

So just as an example, values are not given any thought here.. So it is better to use two steel tubes ( square ) to make up the sides of the frame instead of using a steel plate for the same purpose ? Attached a paint picture to illustrate so there is no confusion.


Won't it be more difficult to get this perfectly 90deg with welded tubes rather than using bolted plates ?

[wizard]

Could you please elaborate on this ? Just so I'm getting it correctly ( English not my primary language ).

I will give this a shot, if you don't follow my explanation then post back with more questions.

The problem with plate steel, sheet metal if you will, is that it needs to be relatively thick to maintain stiffness over a long distance. The classic way to see this is to pick up a 16 ga piece of sheet metal at one end and notice how it bends as you lift it up. Now take a pieces of 16 ga square tubing and pick it up at one end. You will likely see no bending in the square tubing.

Now plate steel is basically the same thing, relative to the machine it is very flexible. To eliminate that flex you have to consider rather thick and heavy sections of steel plate. The plate ends up being so thick and heavy that it dramatically impacts the cost of your machine. In a home shop it can impact your very ability to put the machine together.

So just as an example, values are not given any thought here.. So it is better to use two steel tubes ( square ) to make up the sides of the frame instead of using a steel plate for the same purpose ? Attached a paint picture to illustrate so there is no confusion.

Generally yes it is better to make use of square or rectangular tubing for a machine frame. If you look at the various build threads you will see many different approaches to machine frames built from tubing. The key here is that the goal in each case is a stiff chassis upon which to build the rest of the machine.

Won't it be more difficult to get this perfectly 90deg with welded tubes rather than using bolted plates ?

In each case the prep work on the steel has to be done with considerable precision. In the home shop I actually think it is easier to square up steel tubing rather than plate for welding. A steel tube might only be a few millimeters thick so if you cut it as square as possible you can further fine tune the squareness with a 4" grinder and a fine sand paper wheel. It can be tedious to square up a tube in a home shop but it would be far easier than trying to square up one inch thick plate. Take your time though and you will get the end as square as your measuring device.

That being said if you have a bunch of tubes that need to be square and of the same length it is probably advantageous to find a machine shop with a horizontal mill that can square the tubes up and keep the lengths consistent. In the end it isn't easy to build these machines with out some access to machining equipment. It is doable with very basic tools but it takes a long time to square steel tubing by hand and get more than a couple to the same length. As noted I have used a 4" grinder in the past but if you really wanted to you could use a file (German milling machine) and lots of muscle.

[oveaune]

Thanks for the clarifications, helped alot.

I think I might try getting a professional welder to weld the frame just to get it correct. Will it be required to mill the top of the frame to ensure that it is flat and both sides are exactly the same ?

Another question What accuracy should I aim for when it comes to ballscrews and linear guideways ? This maybe comes down to how accurate and stiff the frame is ?

[wizard]

Thanks for the clarifications, helped alot.

I think I might try getting a professional welder to weld the frame just to get it correct. Will it be required to mill the top of the frame to ensure that it is flat and both sides are exactly the same ?

Welding by its nature induces distortion and stress in a frame, how much distortion is hard to estimate. One approach is to simply minimize welding on the beam the rails will set upon. Even so of many uses these steel or aluminum beams won't be flat enough so in the end machine work is required. This is why some extrusion suppliers offer a machining services for their beams. In the end what you are trying to achieve machine wise dictates what length you go to. Remember many people build gantries out of wood and get good results.

As to your question about milling the rail mounting surfaces I would say this, you will likely need to do something to get a flat mounting surface. Milling might be the easiest and at times the cheapest solution. Some have used epoxy leveling techniques, hand grinding and scrapping or even shimming of the linear rails. In my mind a machined flat surface beats these approaches assuming you can get it done locally at a decent price.

Another question What accuracy should I aim for when it comes to ballscrews and linear guideways ?

Again this totally depends upon what you are trying to achieve. A guy doing PCB routing might need good precision and excellent repeatability, a guy doing home improvement work in wood might be happy with ten thousands of an inch over a foot. The same guy trying to build a wood clock might be frustrated by anything more than one thousands over a foot. Generally work in metal and composites is usually fairly precise.

In other words you need to fit the hardware to your intended usage. That and economics are the gating factors.

This maybe comes down to how accurate and stiff the frame is ?

True; the motion components won't do much good, even if they are high precision, if the rest of the machine can't support them. The whole system has to work together to deliver acceptable results. The material you listed in your first post implies a wide range of materials and that possibly precision will be important for you. If so you will need a more robust machine than one dedicated to wood working.

[oveaune]

Understood, well thanks for all the help wizard, I probably will have more questions as the design progresses, but I think I might have enough to get the frame design started.

Hopefully I can get simulate the framedesign in solidworks ( not sure how yet but... ), should give me a good pointer on the stiffness ( simulating deflection ) of the design,

[cpr]

Question from a newbie: What exactly is meant with a "milled rail mounting surface"?
Are we talking about "planing" the rail, or the frame?

What I mean is:
- is it "enough" to mill at a certain z depth along one axis,
- or do I need to mill along a rectangular path?

[oveaune]

When the frame is welded together, there is no ensurance that the two sides the guiderails / guideways are mounted are at the same height. By milling the mounting surface you can ensure that the gantry has a perfectly level surface to slide on.

Otherwise you might get different height depending on which side of the gantry the spindel is. If you exaggerate it, you might look at it as a big triangle, where the hyp is the movement axis of the gantry.

[hanermo]

Its a good thread.
I built my milling machine in steel-and this was the right choice.
My machine is 2000 kg in mass.

Linear rails are shimmed on y, or on the bridge x-axis i made a grinding sled.
I use a quill (bridgeport) for z.

Alu and steel are about equal in stffness per mass.
Alu has 3x the thermal expansion.
Alu is 3x the cost per mass.
(Iow - alu is 3x worse for the cost, but looks nice. Nice looks wont make it flat, and wont make it rigid).

So alu is about 3x more expensive for the same rigidity, and has 3x the thermal distortion.

[oveaune]

My machine is 2000 kg in mass.

Wow 2000kg How big is it ? I've did a rough weight calculation for my intended size and did not get anywhere near this weight

Did you use hollow or solid profiles ?

[wizard]

Question from a newbie: What exactly is meant with a "milled rail mounting surface"?
Are we talking about "planing" the rail, or the frame?

Planning the mounting surface could be done, that is a different operation than milling. You could have the unit ground too. The idea is to get a flat surface to mount the rail on. This to overcome any welding or manufacturing distortion in the beam. Steel tubing isn't perfect from the factory so you potentially have irregularities that need to be removed.

Depending upon frame size you can also get the two rails in the same plane and parallel when the mounting surfaces are ground, milled or whatever. . This is nice but obviously you run into problems if the frame is too large for your local machine shop. Often you can flatten each rail and them get them coplanar and parallel at install time for a really large machine. I don't think many have to worry about such large machines in this forum.

What I mean is:
- is it "enough" to mill at a certain z depth along one axis,

Well ideally the beam would be flat and next to nothing would be removed.

If the beam is excessively warped then you run into a problem where too much steel has to be removed. It pays to inspect before you buy because the journey from the mill to the supplier is often a long one.

- or do I need to mill along a rectangular path?

You need to mill mounting lands for the linear rails of your choice if you need to achieve a high precision in travel over the length of the beam. You must remember some people build perfectly acceptable machines out of wood. If the beam meets your needs you might get away without machining it. The only caution here is that irregularities can cause linear rails to bind which can lead to premature wear.

One other thing that machining can do for you is to provide registration shoulders for the linear rails. This allows you to easily mount the rails parallel to the beam axis and each other. Assuming of course that the beam is properly mounted in the mill.

As for rectangular path, I'm not sure why you would want to do that. All you are concerned about is where the rails mount.