New Machine Build- My Summer Project (CNC Mill)

[Zac.Hester]

I've been lurking for a little while, and gathered a lot of inspiration for a small CNC Mill I've wanted to build for a while. I always backed down from building it because I wasn't sure of the results, if it would work, and how much money I would spend before I would know if it would work. Seeing a lot of the machines on these forums, as well as one made out of pipe and aluminum U-channels, I decided to take the leap.

Just by way of background, my degree is computer engineering, so my focus might be a little more on the drivers, interfacing, and software, but I'm having a lot of fun designing and building the machine.

My requirements aren't all that great right now (the biggest requirement is the learning involved). So, I really just want something that will make decent PCBs. If everything works out really well, I may be building a second, larger-scale machine with greater attention to precision and range of material.

I've attached one of the more recent renders of my mechanical design. The design is an exercise in keeping costs down while making it easy to build. I have access to a fairly complete machine shop, and a lot of the parts are designed around scrap that's freely available to me. One of the things that might be somewhat unique about my design is my linear guides (ways). I got the idea from an old printer. Each axis is supported by four bearings (made out of PTFE) that runs along two drill rods. This gives me a really simple bearing cross-section that I can reproduce with normal bits on a mill.

I'm also making my own anti-backlash nuts (I have some free blocks of PTFE) and motor couplings; mostly because I'm cheap, and I like having an excuse to work in my friend's machine shop.

Because this forum helped convince me to attempt this large of a project, I thought it might be beneficial to anyone else who's on the fence to see my project's progress, mistakes, successes, and a hint of my design process that years of engineering have instilled in me.

I have a lot of (somewhat unorganized) design documentation on my personal site. I try to keep a log of my progress updated fairly frequently with lots of picture. I will certainly post back to this thread any time I reach another big milestone in construction. For now, I've attached some photos of the first bits of construction.

Thanks CNCzone community!

[cncbob]

I like your concept. Should be a cool build!

[Jesse B]

Very nice concept so far. One thing that is bothering me are the linear rails. First, they appear to be very small in that rendering; they need to be quite a bit larger than that. Second, they appear to be unsupported; I would highly recommend going with supported rails to eliminate any flex. If you go supported the rails can be smaller than if you go unsupported. Wanting to make PCB's, remember that you're trying to shave a small fraction of an inch of material off, so any rail flex could potential ruin the board.

[Zac.Hester]

Originally Posted by Jesse B Very nice concept so far. One thing that is bothering me are the linear rails. First, they appear to be very small in that rendering; they need to be quite a bit larger than that. Second, they appear to be unsupported; I would highly recommend going with supported rails to eliminate any flex. If you go supported the rails can be smaller than if you go unsupported. Wanting to make PCB's, remember that you're trying to shave a small fraction of an inch of material off, so any rail flex could potential ruin the board.

The render doesn't show it very well, but each of the rails is supported at the halfway point of their span with a smallish bracket (since the bearings never have to traverse the exact midpoint). I bought some drill rod (one of my first purchases) to see what kind of flex I can expect over 12" of travel (the longest span any of them go unsupported). Admittedly, I'm not 100% sold on the current size (3/16"). Since I haven't built anything that relies on that diameter yet, I can just bore out the bearings and the mounts a little more, and I can go up to as much as 1/2" diameter rods. I've even considered chrome-plated rods (similar to what you find in printers and scanners). I think the price of the 1/4" chrome-plated rods on McMaster-Carr were somewhat dissuasive. But, intuitively, I think you're right about it being a weakness in the current design.

Thanks for the tip!

[Jesse B]

I'd go up to the largest size you can if you are set on going that route. I hope it works out for you Supported rails are nice, but there is a price involved. You'll soon learn (if you haven't already) that CNC Machines are a very expensive hobby. I'm around 5x what I expected at first for my budget so far

[metalworkz]

Hello Zac,
I agree with Jesse's comment about the linear rails appearing to be on the weak side, and the need to try and make them fully supported does seem to make sense. I also think the upright tubing that supports the Z and spindle motor may need some gussets in the corners, and possibly extending the framework to make it more of a bridge construction to each of the front corners may make it even stronger. I would at least look into adding some gussets to strengthen the Z as you won't want that flexing any more than the rails. Interesting build idea, keep us posted on your progress.

Regards,

[Zac.Hester]

I've decided to upgrade to 3/8" shafts. For now, I'm going to buy some tool steel rods. If those turn out to have some flex (when applying 5ish pounds of force across 12 inches of span), I'll go up to some heavy, chrome-plated rods.

Originally Posted by metalworkz Hello Zac, I also think the upright tubing that supports the Z and spindle motor may need some gussets in the corners, and possibly extending the framework to make it more of a bridge construction to each of the front corners may make it even stronger. I would at least look into adding some gussets to strengthen the Z as you won't want that flexing any more than the rails. Interesting build idea, keep us posted on your progress. Regards,

Actually, my first frame design was to use gusset plates in both the top and bottom corners of the Z-axis tubing. I have enough clearance to still fit some gussets if the need arises. So far, that part of the frame (intuitively) appears to have the necessary rigidity. For novice arc welds, they turned out a lot stronger than I expected.

I've also considered the "full bridge" idea. Indeed, that may be on my horizon for version 2--especially if I like this tool enough to want to mill harder material.

--

The forums JPEGinate my crisp renders. This might look better:

http://planetzac.com/media/2010/04/original/mill_05.png

[FandZ]

+1 for more Z axis support and thicker rods.

Also brass bushings are cheap and very effective. I don't know how much force is generated when you mill pcb's but having the capability to mill harder materials sure is nice. I don't think it'll take too much effort to build a stronger frame. With you being able to weld, it really opens up some options for you. I'm on my second machine and there will surely be a third. If I could go back I would have a built a stronger 1st machine. That's the one thing everyone realizes after their first machine.

[Zac.Hester]

The Z-axis support is making me rethink a few things (I should have posted initial designs here before I started cutting tubes--you guys are awesome).

Here's a thought: inside welds on gusset plates are a little trickier for a newbie welder like myself (one of the reasons I backed down from the gussets). However, I'm thinking of cutting some tubes at 45 degrees on both sides so as to just barely remove the inside wall. Then, I can very easily edge weld those as a sort of dual-action gusset. (I'll draw up a picture and post it here tomorrow morning.) I'm not an M.E., so I don't know if the whole idea behind a gusset plate is that the force is applied to the center of the member it's supporting. In my case, the resistance would be mostly at the edges of the members. In your guys' experience, would the advantage of a real plate be so overwhelmingly better, that my idea isn't worth the easier fabrication?

All of your comments have given me a lot to think about. I'm already sketching out a few new frame ideas for machine #2.

One of my requirements (that I haven't expressed here), is that I can take apart most of the final machine for storage and moving. This aspect will allow me to come back to just about any part of the machine and modify it, or just scrap it and make a new part with the same overall interface.

I think I'm going to recklessly push forward on my 3/8" tool steel guides because that's really the thing I want to find a fault with. If the rest of the machine performs as expected, and the guides fail miserably, I'll be just as happy (and McMaster-Carr will be even happier with the sale of a bunch of proper linear motion rails/bearings). If there's one thing I want to take ownership of in the mechanical design, it's those probably-shouldn't-work-but-just-maybe rails.

I think one thing that might be causing the unanimous push for heavier parts is that the scale of the machine isn't apparent in the images I've supplied. Just for comparison sake, the primary frame components are made from 1.5" tubes. The vertical tubes supporting the Z-axis are 11 inches tall. The tubes that boom out to the Z-axis extend 5 inches. The estimated travel is 12" x 7" x 4". Certainly, there is plenty of room for problems due to flex, but just looking at some of the parts on my bench, I feel pretty good about their rigidity for cutting copper foil with a Dremel tool, and even some patient work with aluminum parts. I'm going to work on getting some pictures for size comparison.

Seriously, though, thanks for all the input. I will use it all one way or another.

[RomanLini]

Originally Posted by Zac.Hester I've decided to upgrade to 3/8" shafts. For now, I'm going to buy some tool steel rods. If those turn out to have some flex (when applying 5ish pounds of force across 12 inches of span), I'll go up to some heavy, chrome-plated rods. ...

Even 3/8" is still very thin. 1/2" is more like it. You can buy 304 or 316 stainless steel rod in 1/2" or even 5/8" fairly cheap from a metal suppliers, they usually have it in polished finish too. If it's only like $20 difference in total price I would go for the thickest rod the budget allows.

Supporting the axes at the half way point is good, you get many times the stiffness that way. Also if you clamp the rod tight in a decent thickness piece of material (on all 3 mount points) it will be considerably stiffer than using thin material and a bolt in each end of the rod.

I know it *seems* that 3/8" will be plenty rigid for fine cutting jobs, but once it is cutting you will almost always be looking for ways to make things more rigid. The problem is not flexion outright, but the vibration caused by the cutting tool will make the rails resonate like guitar strings. That causes bad cutting finish, blunted tools, extra wear, missed stepper steps, nasty burrs on your PCB etc etc. It's really worth the few extra dollars to beef up those rails!

[metalworkz]

Hello,
I think the gussets could be added with or without welding. The placement of the gussets can be altered to enable adding them with a couple of dowels and some bolts threaded into tapped holes in the tubing. You could cut the gussets and use them to transfer the hole locations to the tubing and add them to the outside surfaces of the tubing. If the inside clearance of the machine throat is of concern you could also fabricate the gussets so as to keep the inside throat clear and make them in an 'L' shape or modified 'L' shape to gain more strength. Actually you could very easily make the gussets with holes(say .187" dia to .25"dia) to use as plug weld holes and then clamp them to the tubing and simply fill the plug weld holes and grind flush when done.
I just mentioned the dowels and bolts as a method for being able to dis-assemble things later on. There are others methods to achieve the same results with the gussets if you modify your design accordingly.
You may be right about the strength in your design, but it will not take much to build some more into it at this stage whereas it can be very disappointing to find out too late and have to rebuild sections of the machine or the whole thing.
Once it is functional you will be surprised at how many other things you will find to make that were not part of your original design considerations, so you may be better off to overbuild a bit. Keep at it, it is going to work out.
Regards,

[Zac.Hester]

Originally Posted by RomanLini Even 3/8" is still very thin. 1/2" is more like it. You can buy 304 or 316 stainless steel rod in 1/2" or even 5/8" fairly cheap from a metal suppliers, they usually have it in polished finish too. If it's only like $20 difference in total price I would go for the thickest rod the budget allows. Supporting the axes at the half way point is good, you get many times the stiffness that way. Also if you clamp the rod tight in a decent thickness piece of material (on all 3 mount points) it will be considerably stiffer than using thin material and a bolt in each end of the rod. I know it *seems* that 3/8" will be plenty rigid for fine cutting jobs, but once it is cutting you will almost always be looking for ways to make things more rigid. The problem is not flexion outright, but the vibration caused by the cutting tool will make the rails resonate like guitar strings. That causes bad cutting finish, blunted tools, extra wear, missed stepper steps, nasty burrs on your PCB etc etc. It's really worth the few extra dollars to beef up those rails!

I guess I wasn't considering the more subtle problems from a rigid structure such as resonance. I think you sold me on the 1/2" rods just because of that.

I can't figure out how to get a link out of McMaster's web site, but here's the specs on a bar I think I like:

Material
Multipurpose Stainless Steel (Type 304)
Finish/Coating
Unpolished (Mill)
Shape
Rods and Discs
Diameter
1/2"
Diameter Tolerance
±.002"
Length
36"
Condition/Temper
Annealed
Hardness
140-223 Brinell
Yield Strength
30,000 to 35,000 psi

(Slightly abridged.)

I can go up to 6-foot lengths and get "mirror-like" finish, but the price jumps considerably. Would the unpolished version be smooth enough for Teflon or brass bearings? Or is that finish similar to something like hot-rolled steel? (The only thing I know about stainless steel is that they make a lot of kitchen equipment out of it, and it's tricky to weld properly. Oh, and DeLoreans are cool.)

I looked at 1/2" chrome-plated steel, and I might as well just buy proper linear guides and bearing trucks for that price.

I have some stronger steel angles I can use for the brackets supporting the rods, and I was planning on securing them with shaft collars.

As for general resonance dampening, I have a little experience working with composite structures (mostly in aerial robotics). I might have a few tricks to control those problems later in construction (breaking up standing wave resonance, and de-tuning mechanical members). To a E.E., all these mechanical parts look like resonant circuits.

Finally, what's your guys' opinion on these:

http://www.smallparts.com/Precision-..._rd_s=center-3

(Holy giant-link!)

Do you think I would be sacrificing some rigidity if I use plastic shims? It almost seems like they'd make good vibration dampeners, keeping the oscillations from transmitting (very well) between the sub-frame parts and the frame. And, since I'll probably need varying numbers and thicknesses at different joints, it would eliminate a lot of harmonic resonance.