Building a CNC Mill from scratch. Without castings!

[nerobro]

This is an adventure that started some four years ago. I love to make things. I realized a lot of the things I want to make, need real tools. We're not talking the toolchest in the garage type tools, but the sort of tools you see in 1800's photographs, with big wide leather belts driving machines the size of 1950's caddys, and men missing fingers.

So, I started collecting tools. I got a good hacksaw. Then a drill press. Then a metal cutting band saw. And a bench grinder. But two things always stood out. I needed a mill, or at least a shaper. And I needed a lathe.

While scouring the internet, I found a site run by a very cranky old man. His name is John C Kleinbauer. He runs www.crankorgan.com. A website straight out of 1996. Lets just say it took me some time before I felt comfortable sending him money for plans. That said, money was sent, plans did come.

The plans were for the Brute. http://www.crankorgan.com/brute.htm A 3axis cnc rig that with proper setup can do .003 repeatably. John says it can handle light aluminum work. Which is all I'd expect it to do.

I made a couple orders to mcmaster carr. Ebay provided me with the PVC sheet, and stepper motors. I had a friend of mine go and make a run to a window company to get the aluminum rail. She got it in some excessive lengths, so I'm going to go against johns recommendation and expand the mill a bit.

More or less, that's where it ended. I collected most of the materials, then life just got busy. All of my motorcycle projects are as caught up as I want them to be. I have time some evenings to put in manual labor... So the mill has seen progress. Serious progress.

In the last week, I have laid out most of the parts. Roughly cut them out and even finished a few of them.

The first night I tried just drawing out the parts, with intent of gluing them down to the plastic then cutting them out. Because the plastic isn't 8.5x11, and I didn't plan around that... well that plan didn't work so well. So plan B, was gluing the graph paper to the plastic, drawing out the parts, then cutting them out. That worked a lot better. I still have a couple to go, so I will get you pictures of this.

I have a metal cutting bandsaw. Metal cutting bandsaws cut slowly, and have big wheels to suck the heat out of the blade before it makes it way back into the cut. This also means they cut plastic quite well.

The only cleanup I've done on these is to run their edges against each other to knock the burr off the back side of the cut. This is 1/4" grey pvc.


So, the black lines are the actual shapes I want. To take the cut right up to the black line I'm using a hand file. A 12" fine crosscut file. The first night I cut out the first table, and filed it, on the edge of that table. The height was wrong, the clamping solution sucked. And I ended up with a kinda nasty shaped part. I'll get pictures of that later.

Second night of filing, I came up with this.


Top of a stepladder, big C clamp holding down some 5/8" aluminum plate, and vice grips holding the working piece, cushioned by some strip aluminum. This lets me hold the file vertically, and get a nice cut.

This isn't the best edge I've done, but it shows you some idea of the sort of accuracy you can get with a file. Yeah, I rounded the edge a little. Shame on me.


Finally we have the results of my second days work.


I learned a lot. First, keep moving. If the PVC gets even a little warm, it doesn't file very well. And it's a lesson i had to re-learn later. Second, keep the pressure high, the file needs some force for it to bite.

My drill press was still in storage at that point, so I didn't go about drilling and tapping everything. A couple later I went and got said drill press. I drilled and tapped the first (nasty.. remember..) table. And the two legs that go under it. On my third night of work, I actually had a complete table to slide in a rail. Sadly, I had no gibs, so it didn't fit very well.

I made a run to a couple stores to fix that problem. When I made my mcmaster orders, I did not buy #4 washers. So I bought those. I also bought some #10 wood screws. Which were exactly the wrong screw to buy to hold the end plates on. Turns out I needed #10 sheet metal screws, and even that might not be necessary. I think I'm just going to tap out the aluminum channel for 10-32 threads and go from there.

The second stop was target, to buy a cutting board. Cutting boards are HDPE (usually) and are an ideal material for the gibs. $12 later I had three cutting boards, two of which will go in the kitchen, one will get sacrificed to this project.

[aSpiffyMofo]

Awesome project. I'm looking forward to seeing your progress

[behemoth]

I think i'll keep an eye on this thread.

[nerobro]

I'm reposting this from somewhere awful. it is my project... I actually am more or less "finished" with it.

[nerobro]

Oct 19 2009

The majority of the frame is aluminum. PVC's inability to transmit heat is a problem though. Not one that will be an issue in this project. :-) Thankfully. I think I paid $20 for all of my PVC sheet, fiberglass sheet would have been much more expensive.

And have you ever tried to tap fiberglass? I'd be paying out the butt for taps. and resharpening drill bits all the time. :-/ The only way to effectively drill fiberglass requires carbide drill bits. I have made a few circuit boards in my time, and at $5 a pop, drill bits get expensive. If I wanted something stronger than pvc, I'd be looking at using aluminum. (and I just might... this thing has a table big enough to re-make most of the parts in aluminum later.)

So... pictures.

Here's the table the night I finished it. I got all excited and ran up to my bedroom and took this shot of it sitting in the rail. The holes drilled in the ends of the gib holders, and the edges of the table are to hold the guides/gibs that provide a way to locate the table on the rail.


Oh how nasty that job is. You can see where I screwed up. there's two sets of holes in the middle because I drilled clear through the table with the countersink. I'll use this for the Z axis, so the asymmetrical drive may actually be an advantage. In the lower left corner, you can see where I screwed up on one of the gib mount screws too. whoops. I just re-drilled that hole 1/4 inch in, and called it a day.


And a clear end view of a table. There will be three of these. Since I chose to go with longer rails than the plans called for, i'm going to need to make the tables larger for the X and Y axises. That will make it more rigid as well. I need to sit down and do the math on those axises.


That really is commercial window frame aluminum extrusion.



It gives a great profile for mounting the tables.

I'll likely have the gibs done, and the table fully attached to the rail tonight, or tomorow. Pictures will follow.

Since it hasn't been asked. Total cutting area will be a space roughly 10x7x6"

[nerobro]

Oct 20 - 2009

So I was all kinds of wrong on the cutting area. This thing should be able to cut 10x6x6. That's signifigant given the sorts of materials going into the construction of this thing.

So... without further ado:

I finished cutting out gibs. There are eight per table. one at each corner, and two at each end holding the table into the rails.


If there is any chance. Any chance at all, that the drill you're using in a drill press might grab the item you're drilling away from you, clamp it down. Or in this case, hold it with something that won't let it take the peice away from you.


This is also bad practice. But I am an expert... at least I've not lost any body parts to machine tools. I was trying to get a picture of the white furball that was around my drill bit. Evidently the shutter speed is fast enough to freeze frame a 3200rpm spindle. That ball of white stuff, is actually spinning. If this were metal, I would have stopped and broken the chips off.


Deviating from the plans I bought, I decided to give tapping the rails a shot. The "usual" cutting fluid for aluminum is wd40. I didn't feel like going to bed smelling like light fractions of petroleum, so I went with 10w40 motor oil. It's almost never a good idea to cut metal dry. IIRC the exception is cast iron, but that has it's own set of nasty crap to deal with. I did break a rule while tapping these holes. You're supposed to back off 1/4 turn every full turn to break the chip that the tap is cutting. I didn't do that here... every 3/4 turn the metal stops.


See, screws fit! And wow do they hold well, despite only having 3/4 a hole to thread into.


Speaking of screws. Here's fitting the first gib onto the table.


And now fitting the first table retention gib. There are two of those on each end of the table. You can see I have both side gibs in place.


The table is complete, and one of the endplates is installed.


That's where we end this saga tonight. I need to make the secondary retention screws for both end plates, and cut out the hole for the lead screw, lead screw thrust bearing, and eventually a place for the stepper motor to mount.



Forgot a picture from last night.

[nerobro]

oct 22 - 2009

I put in an hour last night. I drilled the main mounting holes for the second end plate for the Z axis. I also drilled holes in the end plates for the leadscrew. This actually looks like something useful now.

I also did some work on the "big" table. I finished up the top and bottom edge of the gib rail you see clamped there. I still need to clean up it's ends. I also finished the edges on the big table.


oct 23 -2009

Today I got a package from mcmaster. I think that means I got my order in two days. I'm more or less in shock on that. I still am up in the air over what bearings i'm going to use. I have some roller blade bearings around here somewhere that I think will be my final answer.

Tonight I drilled the big table. I finished filing the other gib rail for the big table. I also took some time to think about how to do this well. So I ended up setting up jigs for drilling everything. That let me drill the gib rails very fast. You'll notice all the holes on the big table are clean and straight. When I only need to concern myself with lineing up one axis, I can get things very much on the money.

The big reason for the order, was for shorter 4-40 screws. Look, they're flush!


Lets compare tables. The small and dirty one is 4.5x5.5", the big one is 8.5x5.5" If you look closely you can see one row of holes is populated with screws already. I had time to tap, and install one of the gibs tonight. I am very proud of how straight those rows of holes are.


The installed gib rail, and the unfinished gib rail. The one that's laying down has 13 holes that need tapping.


I thought I might show you the difference between the two axises. The small axis has a 12" extrusion, and the big axis has a 20" extrusion.


All of the sudden this is starting to build itself really quickly. This may not take a couple weeks more like I thought it would.

[nerobro]

oct 24 - 2009

Yet another pile of parts. This time end plates, and stock to make gibs out of. Sadly those gibs didn't get made tonight.


What did get made were the drive plates for the tables. Here I am preparing to drill the hole that will then be tapped 1/4-20. This is a severely undersized drill bit. IIRC it is a 5/32 bit.


And after we drill, then we tap. HDPE taps much better than PVC. And it never showed signs of melting no matter how slow I tried drilling.


And the drive block mounted to the table. The plans call for a floating nut here, but I don't think I'll need that in my case. As i found out later, my confidence wasn't misplaced.


Here's how I made sure the holes in both end plates were in the same spot. I used the mounting screws to align the end plates when I drilled the leadscrew hole. That worked really well.


Speaking of lead screws. On the left is a roller blade bearing. On the right is three feet of threaded rod. Or a lead screw, as that is it's intended purpose. The lead screw is 1/4-20. That bearing has a 7mm hole down the middle. 7mm is a good bit bigger than 1/4" That means the fit will be really sloppy. I don't want any slop, backlash is the devil, and if the rod isn't centered, it will whip and bind. Binding is only second in evil to lash.


The fix to that slop comes from my electrical box. Some copper wire wound around the leadscrew, and you suddenly have a nice, soft, spacer.


Tighten some nuts around that spacer, and you now have a slop free fit. These are abec 5 bearings, so they don't have much play in them. But I don't want ANY. The solution there is to tension them. The design, as per plans, just has one bearing at one end, and the other end other end of the leadscrew is held up by the stepper motor.


I didn't get shots of drilling out the end plates to hold the retaining screws. But here's the bearing mounted to the endplate. You'll notice there's a washer under, and over the bearing. The inner race of the bearing is a little wider than the outer race. The washer works as a spacer.


And, it's hard to tell, but I put one bearing at each end. So I can tension the whole lot.


Oh yeah, it works too. .... wow, I sound like a dork.

http://www.youtube.com/watch?v=PL9EuQsRSmw

So far, my deviations from plans are:
4" more travel on the x axis.
2" more travel on the z axis.
Dual bearing, tensioned leadscrews.
No floating nuts.
Use of machine screws instead of self tapping screws on endplates.
Better bearing retention hardware

I'm sure I'll screw other stuff up too.

[nerobro]

Oct 26 2009

Another day, a little more progress.

On the way home form work I bought some galvanized pipe. I bought 1.5" stuff. The plans call for 1-1/4. But I think I might go out and buy 2" material and completely disregard what the plans say. Bigger tubing is stiffer. And given that I'm going with a longer travel Z axis, some additional stiffness would be good.


Some visualization always helps get a project moving. Vice grips to the rescue. I think I'm going to sacrifice some of the vertical travel, and make a bigger table for the vertical axis. Since that is going to see mostly side loading, a bigger table will make it wear less, and what wear there is, will cause less misalignment at the cutting head. That said, what I have there, will work pleanty fine to get the mill up and running, so I can use the mill to cut better parts for the mill.


Here's my setup for drilling the gib rails. What you can't see is I didn't get it quite done right. so all the holes on the two gibs I drilled tonight are offset. That's ok though, being precisely in the middle doesn't matter much there.


Would you look at that, a pile of tables.


Ever have one of those moments where you go "what the hell am I doing?" Tonight I had one of those. I've had it happen before while building model airplanes. That feeling you get the first time you assemble the separate sections of a wing, and suddenly you're standing next to 7' of wing? Well I put some gibs on the corners of the big table. And decided to stack the parts up to get some idea of how big this thing will be. That was a nice smack of reality...


This is going to be all kinds of awesome when i'm done. I just hope it can cut as well as it looks.

[nerobro]

oct 29 2009

Hey, that's a complete Y axis. I even made a little handle to put on the end to make turning the leadscrew easier. To do it "right" i went out and bought a large drill bit, so I could put a nut on both sides of the bearing. I'm a little anoyed. The Y axis is rather stiff. It turns out that my aluminum for that axis is a little off kilter, and I may need to make new gibs to compensate for it. It's "consistently" off kilter, so I just need to adjust around it. Also the peice of threaded rod, is bent. I have another, it just means going in and taking it all apart again.


I also went out and bought a few more pieces of galvanized pipe. This is the proper height for the Z axis. I am still seriously debating going with 2".


That's really a piss poor picture of the z axis handle. I swear, it's there, and it works well.


I keep trudging along. This will get completed. As long as we're not talking the electronics, I'll bet this will be done in the next two weeks.

Now I am to the point of needing to make a decicion. How am I going to join the X-Y table to the neck and Z axis. The plans call for MDF. I don't like this. My first thought was 5/8" aluminum plate. Which is expensive, and could be bent.

What I think I'll be trying to find, is some scrap 3x6" steel tube, or 6x6" steel I-beam. something tells me those will provide a very sturdy base for this.

[nerobro]

oct 30 2009

So, the images before were just staged.

Tonight, before CSI, I sat down and started working on attaching the X and Y axises. I didn't get any images of me doing the layout, and.. really I errored a bit. Those holes look good. When you drill through metal, if the "other side" isn't parallel with the side you're driling from, it can cause the drill bit to walk. Especially if the flutes are sharp. .... my flutes are sharp. Damn. I had to go in an file the holes oval so they fit.


What did it fit on? It fit on the table for the X axis. Now, these holes are drilled properly, and in exactly where i ment them to be. Sadly... because the holes on the channel drifted. *grumbles*


After CSI, I went back to work.

Here's the plate bolted to the rail. You can see the ridges that caused me the headaches. I was worried that I might need to dremmel the ridges down. Or, if I were really smart, I'd buy an endmill, and cut that down properly. I need to buy some endmills anyway.


Well, here's me just showing off. You can see there's handles on both the X, Y, and Z axises now. And that all the leadscrews are in place.


I've included a ruler to show some scale. The X axis can probally do 6.5", but 6" is the design travel.


The Y axis is 10" design, it can likely do 10.5"


That's 6" of clearance. I can lower the Z axis a little, once I trim the leadscrew.


I really need to figure out the mill mount. I'll make some calls at work tomorrow, see if I can hunt down some nice materials for that base.

[nerobro]

Nov 12

So, I'm not taking pictures as frequently. It's all the same stuff. Do layout, cut out on the lines. Drill out holes, tap the ones that need taping. Wash, rinse repeat.

I wanted this as strong as I could so I decided not to countersink these screws. Rigid isn't a strong enough word for this.


And here's the trim router installed. The fit is T I G H T. And those grub screws are only 1/2 turn past finger tight.


So, because the screw heads stick out on the back, I need to elevate this from the Z axis. I also thought it would be a good idea to extend it down towards the X and Y axises.


I need to make the mounting plate and attach it to the y axis.

[nerobro]

Nov 17 2009

I've got my machinists handbook, and my pocket ref. I should be good for figuring feed and rpm. :-)

Moving on...

Well that right there is 15lbs of steel, and like a pound of aluminum. Yey for raw materials. Well sorta raw. The steel will be used more or less as is. Just as a fun little anecdote, the guy who runs the metal shop, has heard of my dad. My dad runs a machine shop, on the opposite end of the chicago region.


The first order of buisness was some layout. Measure twice? maybe trace once? Well you only get to cut once.


Remember the wrong way of drilling a hole? This is much closer to the right way. If you want any sort of precision, you need to use a pilot drill, or a spotting drill. I can't find mine, and it makes me sad. Thankfully this isn't exactly a precision job.


You'll notice the holes are pretty small. These are pilot holes. The middle of a drill bit doesn't really do much cutting. The smaller the drill bit, the smaller the middle, or web, is. You can use this to your advantage. Drill a pilot hole just a little bigger than the web of your intended drill bit, and instead of needing to press like hell to make it cut, it goes through like butter.

Don't forget to put some cutting oil on those drills. If you expect them to last. Again, this is motor oil, not the prefered cutting fluid, but what the hell, it's handy.

So, I didn't get any pictures of me tapping those holes. These were my first serious holes in steel. 1/4-20 tap in steel. Oh yeah, that plate is 1/2" thick cold rolled steel.


And here's the column mounted to the base plate.


And the Z axis mounted to the column. I have some spacers in there, and I will later need to shim them so they stop distorting the Z axis.


Here's the Z axis setup with the base plate.


And here's a mockup of the final rig. I should be able to make chips this week.

[nerobro]

Nov 24 2009

Well, here we have me using my cheater layout method. I didn't get shots of me fixing my bandsaw three times to make it cut this crap. It came out ok in the end.

So, I made a futile effort to keep the chips out of the ways. I need to tear this down and replace the drive blocks and leadscrews regardless, so I'm not "that" worried about crap getting in there. But it was a good effort. I decided my first work would be to square up the table. Oh yeah, if you're using way covers, make sure the bit can't pick them up. I'll let you guess how I know that.


Well, I knew that the bit would leave something to be desired. I think I underestimated how bad it would be side cutting with a spiral saw blade. Look at that. This is the best looking edge. Tonight I'm going to cut 2" off this cutting bit, and try another pass or two. You like the snow of aluminum chips?


Chips in the ways.


And, this is a working mill.


As it stands, I'm embarassed with the results. We'll see if I can do something less disgusting tonight.

[nerobro]

nov 28 2009

Awesome, my webserver shat itself. Time to build a new one... And talk to the guy who i'm mooching colo space off of. :-/ ugh.

So, since that is all wonky, I have a flickr link.

http://www.flickr.com/photos/nerobro/sets/72157622890557222/

So here's my first real pass. this is 5/8" aluminum in a $9 harbor freight drill press vice, that's clamped to the 3003 aluminum table using $1.49 mini c clamps from home depot.


The head hasn't been trammed in yet. So, that explains this...

You can see dimples and divots where I pressed on the Z axis while cutting. Whoops.

The mill seems pretty capable to me. You like?

That was swinging a 3/4" wood router bit at 6061-t6 aluminum.

Sometime after christmas, I'll be buying the control electronics and making this CNC. But now, the hardware is proven.