[Shalecnc]

Ok, after about 4 months of research & planning, I'm ready to start my build. I'll mostly be using this to prototype airplane parts (lots of foam cutting), but also want the flexibility to cut heavier things like MDF mold forms & AL.

I went with a conventional, though heavy raised, moving gantry design, based on the Misumi HFS8 extrusions. Rough dimensions are 61"x38", with about 48"x28" travel.

Details:

• X-axis beams: Misumi HFS8 80x160mm
• Gantry: Misumi HFS8-45, 90x180mm
• Gantry Risers: 1"x12"x24" AL 6061 plate
• Z-Axis: 8" actuator assembly from VeloxCNC (will replace w/own design later)
• Profile Rails: HIWIN LGW20HCZ0H (x-axis); NSK LY20 (y-axis)
• Ballscrews: 2005x1400mm (x-axis); 1605x950mm (y-axis)
• Ballscrew support: Misumi
• Drivers: Gecko 540, 380oz Nema23
• Spindle: 2.2kw (Ugra CNC), VFD: Hitachi X200

Here's a render -- let me know what you think:

[Shalecnc]

one note - the very long, 6" end mill shown in the render will be for milling foam. For harder materials, I'll reposition the grantry lower on the risers.

[Benonymous]

Obviously not a "budget" build! It looks very substantial with the big bracing plates on the corners. You could probably hog out or hollow the gantry uprights to lose some weight. I was surprised how much the 20mm plate for my Z axis weighed in 6061.

All up, it looks like a superb machine for the purpose you intend.

[Shalecnc]

Yeah this is definitely blowing the "budget" I had in mind when I started out... but surprisingly, moving to the much heavier 80x160mm Misumi beams (with precision cut) only upped the cost maybe 300-400 over the the price of lighter 1545 rails from ebay. Considering that's less than 10% of the total price I though that was a good trade for easier construction and much less flex.

The cost is suprisingly in things like the table top (not rendered below), which ends up being about $500, the fasteners (another $500 or so), and ballscrews and profile rails.

The real cost is in the ballscrews, ballscrew supports, and profile rails at about $1,500 and the electronics & spindle -- another $1,500.

[Johnny@MISUMI]

Obviously not a "budget" build! It looks very substantial with the big bracing plates on the corners. You could probably hog out or hollow the gantry uprights to lose some weight. I was surprised how much the 20mm plate for my Z axis weighed in 6061.

All up, it looks like a superb machine for the purpose you intend.

Hey Benonymous,

Did you get your 6061 plate from the Misumi site automationmaterials.com?

[awerby]

It looks sorta strange to me, Shale. It seems like you've used lightweight materials for everything that doesn't move, and heavy ones for the moving parts. Shouldn't it be the other way around? How much does that gantry weigh? Do you really think those little motors will be able to reverse it at speed without losing steps? For cutting foam or wood, speed is important, and I wonder what sort of performance you were expecting. To me, it seems like a servo system would work better than the steppers for this machine, but bigger steppers (and the appropriate drivers) would be justified if you wanted a stepper system. I like the G540 a lot, but it seems underpowered for this beast...

Andrew Werby
www.computersculpture.com

[Shalecnc]

#awerby > Not sure I'd agree with that assessment. The non-moving parts are by no means lightweight. The Misumi HFS8-80160 extrusions that make up the x-axis are actually heavier than the 8020 3060 extrusions that most reserve for the gantry.

My rough calc. for the gantry wieght comes out to about 140lbs total. (Uprights are about 63lbs, Gantry crossbeams another 50lbs, z-axis, spindle, misc is another 27lbs).

I can save additional weight by hollowing out the 1" thick AL uprights to a degree, w/out loosing much rigidity. You think that's too much for 380 oz stepper to handle?

[awerby]

Yes, I do. Unless you're gearing it down a lot, which will kill its speed, that motor will have trouble pushing your 140 lb. gantry around, particularly when it changes directions and you're fighting inertia. While the motor's rated at 380 oz-in, that's when it's holding still. In motion, the torque is less, and it falls off sharply as speed increases.

Think of it this way: the motor, at its strongest, can hold a weight of about 24 lbs wound on a 1" radius pulley without being overpowered. A bigger pulley would give it more leverage, but a smaller one (like your ballscrew) will have less. (here's a site that lays this out clearly: Measuring stepper motors ) How much force do you think it takes to change directions on your 140 lb gantry when it's going at full speed? Look at the torque curve for that motor and see what the torque is at max RPMs, and decide for yourself if you've got enough. Remember, any time it loses a few steps, there goes your part...

Andrew Werby
www.computersculpture.com

[louieatienza]

I had run my 75+lb gantry using one 425in-oz stepper, spinning two screws via timing belt, with a 24v power supply and xylotex drive, and 8-start (1tpi) ACME leadscrews at over 300ipm. I now run two 425in-oz steppers at 48v (still below the voltage that can be pumped to them) and g540 with a 100lb gantry at over 740ipm! And I routinely cut stuff like aluminum to 1" thickness, G10 and other phenoics, and hardwoods, with no problems or loss of position. I have the accel on the gantry at 35 and 100 on the y (using a 380in-oz motor and same screw.)

My guess is with over 5 times the mechanical advantage of my system, with higher efficiency ballscrews, you'll easily move your axes. The problem is the speed at which you intend to move them. The torque curve remains somewhat flat up to about 250-400rpm depending, which would give you about 50-80ipm max feedrate, with not too much headroom for rapids, since like Andrew says, the torque wull drop substantially past the "corner" speed... You might get more speed with a USB controller board such as a SmoothStepper, and read a lot of reports but have seen no videos in-action (will be testing this myself shortly.)

A suggestion would be to switch to a 10mm or 16mm lead, and use larger steppers, higher voltage psu, and g203v. Another (and maybe better) alternative is to stick with the 5mm pitch screws and get a servo system from DMM or such. The 400W system would give you about 360ipm continuous (feedrate) and up to 600ipm rapids or you could still use higher lead screw with the 750W system... Just throwong out other options here...

[Shalecnc]

So I'd considered using two screws for the x-axis, which would theoretically give me twice the torque for moving the gantry. Thoughts on that approach?

I'm not sure how important rapids are to me.I mean faster is better but most of the bigger foam parts I'm making will just be made once (sections of an airplane fuselage plug), with subquesquent parts being smaller.

If I were to go w/the 2 screws option it raises my cost a bit, how viable is it to eliminate the under-table cross-beam? Getting rid of that would let me fully support the x-axis rails, allowing me to go down to lighter extrusions saving $$.
- I've seen other arguments that you need the lower crossbeam to keep that gantry square.

[louieatienza]

I haven't machined a lot of foam, but if you're going to do larger 3D stuff, unless you do everything in one pass (which I doubt) you're going to require quite a bit of positioning, and all that "air time" non-cutting adds up real fast.

If you have two screws the screws themselves will hold the gantry square (along with your homing switches.) If you use servos, some drives can be set to detect a "hard stop" and that can be used to square the gantry.