The Circuit Farm

… well, 2.5D anyway.

This is by far the most complicated design I’ve even attempted on the CNC machine. I learned a lot. In particular, I learned that the machine is *slow*. This carving took about 14 hours to cut out. I had to break it up over 2 days, and went “a little past quiet hour” on the second.

I could speed it up considerably next time with a few configuration tweaks.

In addition to extra speed, I really need to figure out my tool width — the tool marks on this one were really severe, and it took a lot of sanding to get things to where they are. I might try seeing if I can get it to step “offset” on each pass, to randomize where the tool goes. shrug. I’ll ask around.

I found some source art on the ‘net, but had to turn it into a heightmap myself, by hand. The concept is, you take a bitmap, then color it black anywhere you want the machine to cut all the way through, white where you want it to cut not at all, and shades of grey for all the levels in between. It takes a little mind-bending to stop trying to color it with highlights = white and shadows = black… I am really pleased with how the heightmapping (and thus the 3D nature of the backboard) turned out.

Anyway, without further ado, here it is. My first 2.5D cut on the CNC-to-be-named-later, hot off the CNC, and all ready for Christmas delivery! Kristi did a great job with paint and varnish. I really like the “fossil” effect of the brown.

2 down, 2 to go. The next two need to be designed, too. argh. Maybe I will find something cool that is simple to cut… (edit: I’ve decided to deliver 2 great ones rather than 4 mediocre ones. Watch this space for more signs in 2011)

I finally brought all the pieces together today and cut my first “real” project on the CNC machine.

To this point, I was hamstrung by a variety of startup costs, both financial and temporal.

I wanted to really see how accurate the CNC machine could get, so I picked an easy-to-draw, but time-consuming-to-cut piece, with raised lettering and raised graphics.

The piece is about 12″ x 4″, and came out pretty nicely, if I do say so myself:

…and here it is, ready for Christmas delivery! Kristi did the paint-and-bling job. Very cool looking. She says that the raised letters are more difficult to pull off than the recessed ones, but I think the results are nothing short of spectacular.

One down, three to go. (edit: Just 2 signs this Christmas. More on the way as time permits!)

I might re-cut this out of solid wood, too. It would work well on a two-by-four. The sanding of all those little corners is the trickiest part.
This also took almost 2.5 hours of machine time to cut. The machine is… slow. 🙂

I was cutting 1/8″ deep at a time. I could probably go 1/4″. That would speed things considerably. I could probably push the machine to 45ipm instead of 30, also.
And of course, engraving this instead of pocketing it would have meant the job would have been done in minutes instead of hours; that’s a lot of material to remove.

Very, very fun. I should have cut this thing out months ago; I procrastinated. Simple as that. I bought the software licenses in early October, and had the graphics done before that.

Welcome to the world, CNC machine. You’ve earned your first stripe.

I sent off the broken motor and motor controller today. It will be interesting to see how long it takes to get parts back.

I really need a new way of mounting the router, because the current method eats up a lot of usable bit. So I needed a way to cut through 3/4″ ply. I tried using a regular drill bit for the router bit. That didn’t work.

So I put on the 1/2″ router bit, which is the longest one I have, and was able to carve all the way through a piece of 3/4″ ply. The results were cool, but a little sloppy. There’s a lot of rockin’ and rollin’ happening in the axes as things move around. I think that motor balance is probably a big deal.

I need a “real” end mill. Even a 1/4″ one (which is a bit wide for my taste, but would fit the collet of the router) would work. The 1/2″ bit is a real big object.

Then I got tired and sloppy and I left the Z axis too close to its upper limit before I ran a test. Pop! The transmission nut popped loose. That’s really the weakest part of the design. It took a few tries but I think I found an alignment that will hold for some time. The Z axis is working again.

Then I tried the thing again, and nearly ground the Y axis to dust. So I moved away in Y a little, tried again, and Pop! I popped the X transmission nut. I decided that I am too tired to continue tonight, and will leave the repair for next session.

I need limit switches if I am going to be doing tests near 0,0.

I think there’s a way to tell Mach3 not to allow the machine to go negative. But I am not sure what that is.
Anyway, end mills and limit switches. Need to put in an order.

Mellow greetings from my new robot friend.

The “spare” Z axis motor arrived today. A bit of solder and I was ready to test it out.

It took a little bit of time to get the hang of the basis vectors for the machine (where is (0,0,0), and which directions are up, forward, and away?). Once I got that nailed, though, my first test cut, a square, came off without a hitch.

Buoyed by this success, I spent the next couple of hours learning how to cut text. Turns out it’s pretty simple. All the fonts are available, and the toughest part is figuring out the “machining” part of the equation; my first cut was in millimeters, I decided that I wanted to cut with the text rotated 90deg to make best use of my workpiece space, etc.

I also learned that working with really tiny pieces of wood is not as easy as you would think. Clamping in particular is a bit tough. So my first “real” cuts were done in large pieces that I could clamp from the edges of the table, making it important for me to know exactly how much room the Y-gantry had to move around in (or else I’d hit a clamp…).

I settled on 2″ high letters, and the text ended up being about 12″ long. The reason that it’s a shorter version of the classic “Hello, World.” is because all that extra text would have meant more tweaking of the font size to get it to fit on the cuttable area, and it was getting late. So there you go.

Here’s a photo of Mach3, the “control” software, running the G-code for the text.

The G-code was created by the “CAM” software. Normally, I would have had to use all 3 pieces of the software toolchain to get a cut sent to the machine, but Text is such a basic function that in fact, Mach3 itself has a plugin for it, but I couldn’t get it working, so I decided to check whether the CAM software could do it, and it could, so I did it that way.

CamBam is decently powerful. I had to figure out a few pieces of jargon in order to get it to work. I quickly figured out that trying to “trick” it (by putting in a workpiece thickness of 0, for instance) was just making life tougher. So I put in the actual tool width, work thickness, and desired depth, and things went more smoothly from there.

To test out the cuts, I set the “Z zero” point in Mach3 to about 2″ above the table, then ran the G-code. I made sure that it looked like things were moving in the right direction, and tweaked the CamBam settings as appropriate to make things look right before actually cutting the piece.

A side note: when “engraving”, CamBam added an extra G0 Z1.0, which sent the router 1″ *into* the piece, right before doing the rest of the cuts. I assume that was supposed to be a G0 Z0 to “home” the Z axis, but I couldn’t figure out how to make the CAM software do that properly, so I ended up editing the G-code myself. No biggie.

It took about 5 minutes or so to cut the letters out. I am limited to about 1/8″ depth of cut right now because of the router bit that I am using. When I switch to a longer bit, I will be able to cut out pieces, rather than just engrave in the surface.

For everyone who has been following along (and for the people who find this blog later and wonder where the “build” photos are), here, at long last, is a photo of the yet-to-be-named CNC machine:

…and just in case you are into the electronics parts too, here’s a shot of the computer and motor driving gear:

The computer is an old desktop I had lying around (doesn’t take much to run Mach3), you can see the mouse and mousepad under the machine.
The silver box is the 48vDC/7.3A power supply. I have not yet hooked up the E-stop or power switches (you can see them sitting on top of the PC).
The black box with the DB9 connectors is the Gecko G540. It serves as the translator between the step/direction commands from the PC (via the parallel port) and the coil energization sequence of the motors. You can see that there is nothing hooked up to the “Y” controller, and that the Y axis is hooked up to the “A” controller, if you look closely.

I need to build a case for the electronics. In particular, I would like there to be a power switch for the DC power supply. I have been thinking that it would be cool to put the stuff inside the PC case if I can; it has a built-in fan, and seems like there should be enough room. I’ll look into it.

All in all, I’m pretty satisfied with how well things have gone so far. I’m making sawdust!

Seems like I shouldn’t call these posts “CNC ‘Build'” anymore…