[LBSamurai]

Project: Aluminum Eater (Scratch Build) - The early stages

Like so many others around here, I've been lurking a long while. The learning curve and the process that follows seems like a pretty solid pattern, after all the reading and watching other people get their machines together. I knew I would be doing a build thread, but the golden rule of build threads is to have something to show right from post #1. In other communities, it is generally verboten to start posting with concepts in CAD without having hardware pics right up front, but on this site... starting with concepts is the right thing to do.

I have decided (and got the appropriate permission from the committee) to go ahead with a gantry machine build. I don't really want to call it a "router" since I associate that with woodwork, and I've always been a metals guy. It's an all-metal project, and I plan to use it for primarily aluminum, but also for plastics and soft metals like copper and brass. Maybe I'll dabble with some wood decoratives, or if I want a fixture-puzzle challenge I can carve a nice electric bass from exotic tone woods someday, but I'm really more interested in things that I have to deburr, then tap holes in and bolt together. I'd like to tinker with a 4th axis and come up with tube stocks for bolt rifles, or proto some mountain bike parts. This tool is the key to getting the intake manifold and bracketry built that I've designed to put an Eaton M62 blower and MegaSquirt port-injection system on the 1.6L Suzuki motor in my Samurai. I plan to use it for scratch-built PC cases, and water blocks for the nice peltier-cascade water cooling system I dreamed up and have most of the parts for. Oh, yeah- then there's that 70-amp plasma cutter that has been reliable to me for years. The spindle shall be swappable with the straight-head torch I've had secretly stashed since forever- the machine will need a dual-screw X axis to allow for that. In short, it's a tool I've been leading towards building for quite some time.

There is one catch. My workshop is packed away into two cargo containers back in the States. I'm a career G-man, so I've been on an overseas position in Germany for the last two years, and will probably be here a little while longer. The point is, I've got to do this with what I have currently available to me, and that's little more than my general mechanic's tool kit, a cordless drill, Dremel, and soldering iron. It's truly a new layer of complication, when I am used to being able to easily cut, weld, and drill just about anything. It is requiring a new way of thinking, and makes it doubly important that I get the details right the first time. It will be a lot harder for me to change a small mistake when I am specifying cuts and holes to have someone else machine for me. It hurts my pride a little to put in an RFQ for an aluminum bar with a square cut and a few precision holes that I am normally equipped to do myself, but it just isn't possible to do in my present situation. Alas, I want to do this bad enough, so I'm going to make it work one way or another.







So now that you know what I want to do, here's a little laundry list to organize my design plan.

-Overall Size of the System: I want a functional cutting area of 2'x3' under the spindle, with the additional clamp space to work a piece that size. I'm doing all of the design work in metric, but the American in me knows what that looks like and what types of material will fit on that table, so these three dimensions are imperial. The gantry must clear 12" (though I may reduce this to 10") above the base cutting surface. What do I mean by base cutting surface? That is of central importance- but I'll get to it. The footprint is probably going to be a foot beyond that on each horizontal axis, and I'll build a table to make it a bench workstation in the end... with flood coolant and an enclosure.

-Style: Tabletop moving gantry, dual-screw design. Main work surface will be on a solid sill enclosed by the frame base that will allow it to be switched between T-slot and a steel plasma table. A flexible design for a removeable 4th axis will allow the head- and tailstock to be arranged in a variety of orientations. A modular coolant system that can be installed and removed as routine setup work will be a feature planned from the beginning to implement later, along with the bench and cabinet enclosure to finish it off.

-Materials: Extruded aluminum profile for the main beams, ground fixture plate for the gantry arms and table end plates, and the brackets machined to my spec from standard sized extruded aluminum bar stock by anyone willing to bid on the jobs. I may add dampening material to the table main beams by filling them with sand and casting epoxy plugs in place to give it a more solid base.

-Actual Working Dimensions: Design spec so far is based on rail dimensions of 1200mmx800mmx400mm

-Accuracy, Precision, Resolution, Feeds: As good as 5mm-pitch Chinese rolled ballscrews can get me, which from what I can tell is plenty good enough. This limits my rapids to under 200 IPM (probably less in reality) but it suits my application. The original plan was to use 20mm supported rail, which is what I've spent too much time tinkering with in SolidWorks to get the design figured out. Now I've decided to go with THK rails, and will need to re-work the design as I pluck the parts off of eBay. I can adjust the dimensions to suit the rails I find. The way it is working out, I am building in the opposite order from what I envisioned. It's becoming electronics first, then linear rails, then aluminum profile, then source the machined plates, and lastly order ballscrews from China to fit precisely to the frame. Also different from my initial drawings will be the move to 25mm ballscrews, since the price difference at that length from the 20mm is fairly small. I think the fact that I ended up with 7.5A electronics instead of the G540 solution has me going for a much stiffer build.

-Computer: Patiently watching eBay for a used dual-core motherboard, CPU, and 3-4 Gb RAM that will run under Win7. I don't know if I will be constrained to a 32-bit system for Mach3 or any CAM software. I'll also need to consider the compatibility of x64 with PCI-e parallel port cards. I will need the second parallel for the 4th axis, since the X-axis is slaved and will require the normal 4th channel. Shopping around for touchscreens, and will mount the PC and electronics into a 19" rack with some power distribution and appropriate EMI shielding. For software, I'd like to be able to run SolidWorks from the controller PC just for in-the-shop convenience. That won't be a problem with the component range I'm looking in, coupled with the 8800GTX card that I have lying around to go in it.

-Controller/Drivers/Power/Steppers: I found good deals on this so far, and have obtained five Gecko 202's and a PMDX 125 with i/o expansion cards. I ordered a Cosel PBA1000F-48 1000w/48V switching power supply for $160 on eBay, which is a fraction of it's original cost. When these deals came along, I quickly abandoned the G540 plan since I knew I would need the amperage if my focus was going to be aluminum cutting. I've ended up with some stout components. I will buy five of Gecko's 400 oz-in NEMA 23 steppers and set them at 5A in BPP when the time is right. Everything seems to be well matched on paper. I found some SPEAKON 4-pole connectors and shielded 18/4 tray cable, along with the odds and ends for the wiring and hardware installation at Altex, who was nice enough to ship them to me. Don't get me started on the frustration with getting stuff shipped to APO (It's the regular ol' post office, people! Put it in a damn flat-rate box, is that so hard?). I plan to use the SPEAKON panel receptacles for the stepper cables, and catch the the drain wire from the shielded cable through a banana plug to chassis ground in my electronics cabinet... neat huh?

-Linear motion: Got four THK HSR20R blocks so far... Patience.

-Tool: Thinking about a Chinese 2.2kw spindle on a non-Chinese VFD. That's good for fat tools like drills, bores, and face cutters beyond the realm of trim routers. Getting the most from this type of tool will require my best efforts in building rigidity into the machine. On the far other end of the spectrum, I'd like to integrate my Thermal Dynamics plasma torch into the table. I can run any plasma gas and service gas that I want through it, and with it's 1" capacity it is a versatile machine. It can be tuned for cutting just about any type of metal, potentially being a money-maker if I ever decide to produce replicates of some of the designs I've built already, or whatever else comes to mind. No shortage of ideas, that's for sure. :idea:

So that's my introduction. It will be a slow go, but I thought it was time to start sharing and begin some dialogue with you guys. Please, feel free to post your thoughts and feedback! :cheers:

[ger21]

Mach3 can not use parallel ports on a 64 bit OS.

For cutting metals with a design like that, imo, your only chance is to reduce the gantry height from 12" to about 2". It just doesn't look anywhere near rigid enough for what you want to do.

Here's a link to the only router type machine I know of on here that's used almost exclusively for cutting aluminum.
http://www.cnczone.com/forums/diy-cn..._finished.html

[ger21]

And one more thing. A 400 oz motor is a 400 oz motor. Whether you get one that works with a G540, or one that needs a G202, they both have the same torque. You might see a bit higher rpm's out of the 5A motors, and the power supply voltage should be a bit lower, but i wouldn't expect the actual performance to be much different then if you used a G540 package with 381 oz motors.

[LBSamurai]

Thank you for the response, ger21. I see you offering a lot of helpful advice in others, I appreciate your attention to my thread. Your link to the other build thread has value to me, and I think I will revise my gantry supports based on that, with a twist. I'll post drawings as I get to it. I also agree about the z-axis, however I haven't determined the system I will use- I could put it on 35mm rails and have that subcomponent overbuilt- I think that a Fadal VMC has less rail than that, albeit on a much more substantial set of components to which the rail joins. The primary work surface is near the bottom of the XY-frame, making the gantry supports effectively shorter. The clearance between the lowermost Y-rail and the top of the primary work surface is my concern, I'd like to be able to work with awkward pieces rather than just flat plates. I would like to pick this discussion up later in more detail, and present better drawings. For now, consider the drafts as schematics.

While I agree that a 400 oz-in motor is a 400 oz-in motor, I seem to understand that a little differently. A rating is a rating, and this motor is rated at a given voltage and amperage to provide that holding torque. Devalue those parameters, and you devalue the actual holding torque, even though the rating stays the same. If I apply a 48VDC potential to the poles, whether done at 3.5 amps with a G540 or at its rated 5 amps, when static it should be the same, but when a load is applied, the more current available the more able the stepper is to step against it. I'll have to go back and find the formulae to back-calculate that, but that is my understanding of how it works. If I still seem lost, please continue to talk with me about it. I will need to calc that out when I get to toying with feedrate calculators and things.

I had a feeling that the 64-bit OS was a no-go. Thanks for pointing that out, saved me some searching. Also saves me some cash, all I need is a 3GB PC with hyperthreading and an early dual-core do push the limits of the 32-bit platform.

[LBSamurai]

Okay, inspired by hemsworthlad's build, I have come up with a new base frame to handle heavier loads. Gantry is next. Please share your thoughts. Oh, rail is 20mm and the dimensions are still based on what I specified in the introduction.

[LBSamurai]

Made a little headway on the design revision today. While I have little doubt about its rigidity, I am concerned with the weight and cost of this approach.