Overkill DIY Machine Advice

[TerraHawk]

I co-own a small robotics company and have recently found myself ordering more 3D printed parts. Since I have about
20 years of extra robotic components laying around, I decided to see if I could put together a large volume
(600mm x 800 mm x 600 mm) 3D printer. As it turns out, I have the following items:

- six 25 mm diameter by 1.2 meter CNC rails with twelve recirculating ball carrier blocks
- three 25 mm diameter by 10 mm pitch ball screws with anti-backlash ball nuts
- two NEMA 43 stepper motors 12 Nm peak, 2400 rpm max, 4000 ppr built in encoder, closed loop hybrid controllers
- one NEMA 23 setpper motor 2 Nm peak , 1800 rpm max, 4000 ppr build in encoder, with closed loop hybrid controller
- one welded aluminum cubic structure ~1 meter cube on wheels
- 2 kw adjustable DC power supply

Of course, this is extreme overkill. The ball screws can put out 5000 lbs of force and the rails are rated for 2000 mm/sec.
But I could use it for laser cutting and machining in addition to 3D printing.

Even though I have 25 years experience in robotics, I have NO domain knowledge in 3D printing and would appreciate
opinions on the following questions:

1) I do not want to develop my own g-code electronics. The closed loop stepper motor controllers I have accept dir/step
commands which I can tap off just about any existing board. Are any existing boards fast enough to support my
required 160,000 steps per second?

2) If I drive the ball screws at 1:1 for X and Y, my max comfortable speed is 300 mm/sec. That seems OK. My max acceleration
at that max speed is ~5,000 mm/sec^2. Is that sufficient? My single step distance is 2.5 micro meters. Is that OK?

3) I would like to start with a 400mm by 400mm heated bed. I can't seem to find borosilicate glass that large. Is there any other
material that can be substituted? Has anybody tried hard anodized aluminum?

4) There will be essentially no mass limit on the head. I have been reading a lot about how much trouble dual extruders are.
Is there anything I can do with all of this mass margin to accommodate dual or quad extruders?

Thanks for your time and input!

[awerby]

While I applaud your intention to build a machine out of the junk you have lying around (I tend to think that way myself) I'm not sure that what you've got really adds up to a 3D printer. It sounds like a promising start to a CNC mill, though. A 3D printer, unlike a mill, doesn't really encounter much resistance to moving, but it wants to move fast, with the ability to change directions quickly. Generating a lot of inertia by using heavy components like those massive ballscrews will work against you, I'd think. Even the weight of those heavy rails on the driven axis will be a lot to start and stop, especially when you add the weight of whatever you're mounting them to.

NEMA 43 motors are also overkill for a project like this; they typically have very high inductance, which means they need a lot of voltage supplied to achieve their optimum speed. This one http://jssmotor.en.made-in-china.com...h-4-Wires.html , for instance, has 16mH inductance, which means it really wants 128v, which is more than most drivers (especially inexpensive ones) will tolerate. They generate a lot of torque, but you don't need that for a printer. Most printers use NEMA 17 motors for good reasons...

I don't know where you came up with the requirement of 160,000 steps/sec. Most printers can rapid at about 120mm/sec (actual printing is usually slower), but that works out (assuming you have a normal stepper with 200 steps/rev and 8mm pitch screws) to 3,000 steps per second. Even if you go with 10x microstepping to combat resonance issues, that's still just 30,000 microsteps/second. How fast did you expect to run this thing? If you really want your motors to go really fast without loss of torque, you should consider servos instead of steppers. But even a servo would have trouble spinning at 48,000 RPM, which is what 160,000 steps per second times 60 divided by 200 works out to. Maybe you'd want a spindle that went that fast on your mill, but if you attached that to one of those ballscrews and let it rip, get out of the way quick - you'll be launching components of your build through the wall..

I like the idea of building it in a welded aluminum cube, though. I'd say keep reading up on 3D printers - there's a reason they're built the way they are. You might buy a cheap one just to get familiar with the way they work, and then launch into a building project that uses components that are appropriate to the task. Here's a source for the glass: http://www.sandysindustrialglass.com...33-sheet-copy/

Andrew Werby
www.computersculpture.com