From The Ground Up

[Crushmonkey]

Hi again everyone! I've been away for a while, but wanted to come back and start a thread to document my new DIY project and get some input from you folks who have already done this sort of thing. I'll outline the basics here, and then update as necessary. Here's the general idea of what I want to achieve:

Phase 1: design and build metal-frame gear or worm-driven unit for desktop CNC routing applications - metal-frame for durability and ability to be scaled up to larger frame(s), and non-belt drive for precision and longevity.

Moving-spindle system instead of moving-table.

Software is pretty much open, of course expense is an issue. I don't want to have to learn to write GCode line by line, but I don't need it to be able to compute laser-carving in a bazillion axes (axises?), either. If I can generate the file in Corel or Adobe, feed it to the program, and have it cut my material within reasonable tolerances (within 1/32" is pretty loose, considering, but it'd be a starting point).

I've already found a chip/dust-evacuation system that suits my needs, so that part is covered.

Some of the sticking points include vaccum hold-down for sheet material (a ROOTS blower/pump is freakin' spendy), an oil-misting system for cutting sheet metal (we had one on our Multicam and it was awesome) - I could pay some kid to follow the spindle around with a spray-bottle of vegetable oil, but I'd rather figure out an automated dispenser/sprayer, material depth/height gauging (again, the Multicam had a block unit you'd set on the table, let the tool in the spindle make contact with the block to close a circut and tell the computer where the tabletop was, then repeat the process with the material set on the table).

Since I have access to oodles of scrap plex/lexan, I'm considering building a double-enclosure for the finished unit. This way, I can not only keep my work area clean and somewhat less noisy, but perhaps even setup an air-flow system that will keep whatever chip/dust gets away from the head evacuator from getting into gear teeth or worm drives.

Phase 2: Apply finished, de-bugged design to larger-scale table(s). Basically, I want to learn everything I can on the desktop unit and work out all the kinks I can on a smaller (2x2-4) before stepping up to something like a 4x8 or 8x10 unit (all sizes here are actual cutting space). Also, as my budget becomes more flexible, larger tables make more sense.

The desktop will be for learning and play, but the larger tables will be expected to at least cut letter faces and backs out of acrylic and sheet metal for sign applications.

I'd focus on buying a professional unit, but it would defeat the purpose of me knowing every single little nitpicking thing I can about how it's built, how to maintain it, and what I can expect it to do (at least as far as the hardware goes. As far as software goes, I don't mind paying a little extra to have the computer do most of that work for me). Also, like I said before, with DIY, my budget will allow for a multi-phase project such as this.

So - if I didn't already lose you in the above ramble, basically I need input in these specific areas:

Frame/Gantry: Metal-frame as opposed to MDF/wood/etc. I have access to plenty of steel angle, but I don't mind the cost of aluminum frame. I have access to a journeyman welder (steel/aluminum), but don't mind working with a clamped/screwed table (temporary) for a while until I have it all worked out and finalized for welding (freakin' permanent). Footprint is negotiable within reason, assume a 20x40' shop space for the bigger tables, 10x10 for the desktop.

Spindle/Head: I don't mind using a Rotozip or Dremel for the desktop unit, but will require something more robust for the larger tables. A Colombo is prohibitively expensive, but as long as it will perform within reason for the cost and I don't have to replace the dang thing every month.

Motors/controller: I'd rather skip any belt-driven system and move directly to gears or worm, unless someone has input on how a belt-drive would be better for my application than the other options. I don't know yet the pricing here locally to have gears and toothed rails precision-machined, so this is also something I'd like input on. It'd be nice if I don't have to mod the control unit too heavily to work with whatever motors I decide on.

Material hold-down: I'm considering an electro-magnet hold-down for at least the thinner gauges of sheet, but for thicker materials I'm willing to consider most anything that's practical and time-effective. I'd go totally vaccum-hold, but sadly, I've never found such an animal as a cheap ROOTS blower/pump.

Oil mister/sprayer: I got spoiled with this option when we had the Multicam, and noticed that we went through bits faster and messed up more material when we didn't use it. The mess isn't a huge deal, since the table will be cleaned thoroughly every day.

Software: If I can run it with Linux or Windows 2000 or XP, if it will process Adobe or Corel files to scale, if it will cut 2D and supports multi-pass cuts for thicker material - and I don't have to learn Gcode to use it or rob a bank to afford it, then I'll consider it.

Misc: Speed within reason, I don't know what kind of IPM I'm shooting for yet, but I'd like to be able to throw a sheet of plex or .060 aluminum or a couple inches of sign-foam and have it cut some letters or shapes out at least faster than I could cut it myself with a jigsaw. I can always find something to fill the lag time, but I'd rather not have to find busywork since the larger tables will eventually be expected to make money.

I'd rather not have to install three-phase power until or unless I absolutely have to, but it's negotiable.

See anything I missed? More to come as things happen.

[thkoutsidthebox]

Sounds like 8020 aluminium would be a good option for the desktop machine. I haven't used it but other have with good success.

If you go 2x4 you could look at a router spindle in the 2 1/4Hp range possibly. If not then instead of a dremel I'd get a small router with speed control.

You can have your motors drive threaded screws directly with or without a gear system if you dont want to go through a belt, but the size of the motors will generally be governed by the size and weight of the gantry (Not the other way around). I'm not sure what you want to do, if you want to cut thin metal sheet, maybe you should look at a plasma cutter. If you want to mill thicker metal, then fair enough, work away.

Lots of people use Mach3 to control their machine with great success. Its $159. http://www.artsoftcontrols.com/ I use it myself. It works with Windoze, and possibly Linux although I haven't tried Linux. For the generation of G-Code you mentioned you only want to do 2D work, so possibly look at CamBam, NCPlot, or SheetCAM.
http://www.brusselsprout.org/CAMBAM/
http://www.ncplot.com/
http://www.sheetcam.com/
I use VCarvePro for my G-Code generation. Its a great program, and will work great for 2D.
http://www.vectric.com/

Speed will be governed by the thread pitch on your screws, the weight of your gantry all up, and the strength and type of your motors.

I can't see any reason to need 3 phase power for the desktop machine...someone more experienced may see a reason for the larger machine, but again I dont.

[Crushmonkey]

well for the prototype desktop, it's pretty much open, but I'd like 4x8 for the finished working unit.

I was thinking... Why don't I just cannibalize control boards from a gerber vinyl plotter, and then beef up all the actual motors and such? Wouldn't that be cheaper if I did that and then kept the Gerber Omega program I already have? I could be wrong, but a table is just a plotter that thinks it's bigger... In the sense that instead of rollers pushing vinyl around, the gantry and such move over the material. If I used templates that are no bigger than my actual cutting/movement space of the table, it wouldn't be an issue would it?

If I used threaded screws wouldn't the allthread sag in the middle if I went over a certain length? That's why I wanted to look into a gear on the end of the motors running to a linear rail (or whatever it's called, my meaning is a rail with teeth that match up exactly with the teeth of the gear)

Gantry weight wouldn't be like the Multicam we used to have, that thing was hugemongous. I'd want it robust, but not unnecessarily so. I'd want it to move at a decent speed and not shake itself to pieces so yeah I was thinking angle iron for the base and aluminum frame for the rest.