Grand Unified Theories of CNC/MakerSlide devices

[neocalm]

I am a complete newbie just getting ready to dive into my first ORD Bot Hadron build and then flipping through the forum said to myself "Hey, a 2.x Laser Cutter.. that's exactly what I should be building ..." and then thinking to myself "Why isn't there a CNC Mill based on Makerslide?" - then realizing it was a forum away at Shapeoko. (I'll ignore for the moment the CNC Router, Plasma Cutter, Vinyl Cutter/Plotter, and 3D Scanner that would perfect my garage.)

My eyes are probably bigger than my stomach but even if/when I build all three devices I cannot see using them simultaneously.

I am exploring the sharing of easy-to-share parts between the three. I could tolerate about seven minutes of unskilled unplugging/plugging, and a few less minutes of unscrewing/screwing. I am not talking about squaring / skilled replacement of parts sharing so I'm pretty much assured of buying a lot of Makerslide!

This would probably involve some quick-release harnesses and/or specialized mounts/plugs but sharing parts might rationalize one higher quality part (plus this overhead) over three cheaper parts.

I just wanted to make sure as I get started on this - has someone done this already?

Chris

[lasersafe1]

No.

[awesomenesser]

...
This would probably involve some quick-release harnesses and/or specialized mounts/plugs but sharing parts might rationalize one higher quality part (plus this overhead) over three cheaper parts.

I just wanted to make sure as I get started on this - has someone done this already?

Chris

People have talked about making an all in one machine but I do not believe anyone has ever decided it would be a good idea. The problem is that they each require different things.

Laser: very high speed, very low weight gantry, belt drive, sensitive/delicate optics, z axis isnt really needed (at least a motorized one isn't), enclosed/exhaust needed
Mill: very sturdy gantries, ballscrews for the torque, usually fairly slow, need a strong z axis
3d printer: similar to a laser but an accurate z axis is needed

So already I see some issues popping up mainly with the laser. Lasers need to be lightweight and very high speed (engraving) which is pretty much the opposite of a mill (slow and very sturdy). Also lasers require the unit to be enclosed and ventilated, the mill is usually open and can be very dusty (terrible for laser optics). The mill usually moves around the stationary table, where in a laser the table is moved (because the height is only needed to adjust for the focal distance when cutting never during a cut). Since lasers are high speed they use drives like belts, where a mill needs torque so ballscrews are used. A 3d printer mill combo shouldn't take much modification, just unscrew the router/dremel and bolt on a extruder and mount a heated plate (might not be as quick but it should work).

I am not trying to tell you to give up but make sure you put a lot of thought into the design if you are going to go forward with this. Post ideas and people with the different machines might be able to help you fine tune your designs.

Also kep in mind what these different machines do to decide if you even need all three. (mills can do just about everything a laser can do (although just not quite as precise))

[neocalm]

Thanks for the replies. Again, I wanted to leave the hardware/mechatronics alone because of the different design philosophies (that awesomenesser points out well!) and the fact that I have no experience whatsoever with moments of inertia and other mechanical engineering , but what lead me down this path is that my current BOM for the Hadron beyond the MakerSlide:

. Cut aluminum parts $75 +shipping (thanks mikegrundvig!)
. Misumi/McMaster parts - ~$40 + shipping
. 6x NEMA 17 motors shipped $56
. 2x Shaft Couplers shipped $6
. Large PC Power Supply - had lying around
. QU-BD Dual Extruder w/12V heaters shipped $98

So the Azteeg X3 w/ drivers is actually around a third of the BOM cost at ~$176 (not including the heater bed which is specific to the Hadron).

It just struck me whether there might be a way to share the drivers/electronics/supply in other designs in 15 mins or less changeover to reduce BOM costs. I doubt sharing motors or anything below is cost/time efficient.

Also I suppose I might just consider going 'raw' with an LPT driver breakout, and LinuxCNC at the start, too! With desktop motherboards going to $58 (Intel Atom D425KT) my guess is a Linux box with some software to emulate the various functions is the "high road" for this - if I get time to do some programming or can reuse code...

Chris

[TLHarrell]

Interesting concept in theory, sharing at least the controller between different units. I don't really see any reason why you couldn't other than the fact that for each type of machine the motors, limit switches and other paraphernalia would be set up differently. As far as the software stack goes, you could simply swap .ini files around, or use different software for each iteration. You would definitely use the same computer system to drive them all.

I would take a look at doing something like placing the controller electronics into a removable hard drive bay. You simply power down the tool, unlock and pull the driver tray. Go stick it into the tool you want to use. Power that up, swap configs and start your software stack.

You might also be able to include the motor drivers themselves in this. Anything beyond that would be of lower value to your concept.

[r691175002]

Everyone considers building a unified machine at some point but the more you learn the more you realize it is a bad idea.

If you are only looking for a toy/learning tool it is plausible. The only approach I would really recommend is swappable heads on a laser cutter, because a dremel/extruder is much cheaper than a co2 tube.

Having a router/mill driven by mxl timing belts at 17 nema motors is laughable. Not sure how difficult alignment will be for a 3d printer with a 24x18" working area. There is a reason why most 3d printers only have a couple inches of travel.

It is not worth spending money on discrete electronics for each type of machine. Run it all off linuxcnc/mach3 and do everything with g-code. The only controller that adds significant functionality are the laser boards since you get plugins/print drivers. They also cost ~450$ which could buy you an entire printrbot though.

[neocalm]

r691175002 - Yes, I agree. I was more thinking the mechatronics. Also my garage isn't that big so I'm really looking at a hobby router anyway you look at it - not even a 4x4.

TLHarrell - Exactly! The old 3.5" Parallel IDE hard drive bays was exactly what I was thinking - I hadn't counted the amount of I/O but I thought 40 pins should be able to cover it. It's tight for heat dissipation but I suppose that's what a IDE hard drive fan or heatsink is for. And yes, I was considering both drivers and controller only.

For example, I hear a lot of excitement around a Geckodrive G540 4-axis system, but I cannot seem to figure out whether the Gecko drivers are inherently any better than Allegro at the small NEMA 17/23s. I read a lot about "smoothness" and motor linearity on Gecko's web site but don't know if Allegro's are less smooth / linear for the smaller steppers? Certainly there's a lot more cost for a G540, but as I said in the opening post, if I can buy a "universal" controller/driver that adds value all around, then it makes sense to splurge on the common trunk if it doesn't hurt any of the applications.

So I wonder if a G540 is buying anything (plus I'd need a fifth-axis for the second extruder on a Hadron) besides capacity to go to large steppers for the CNC mill or CNC router. And conversely - are Allegros are better than the G540 for smaller NEMA 17/23 steppers?

Dumb guy questions, sorry:

1. What's the mechanical and mechatronics difference between a Hadron and a $50,000 Stratasys? How do things get tighter and more precise for finer resolution - is it the linear bearings, the belts, the motors, the drivers? Same question for laser cutters, routers and mills.

2. I hear a lot about hybrid steppers (closed-loop) and servo motors and know nothing. The hybrid approach sounds interesting if it can use cheaper encoded steppers with servo electronics that could swap out in the bay. Would this buy anything with regards to a better CNC product?

Chris

[neocalm]

Finally found some other efforts towards a unified mechatronics module:

http://dallasmakerspace.org/wiki/Portable_CNC_Cart

Looks like the MESA 5i25+7i76+ Stepper drivers (looks like the Gecko or Keling DSP would work) might be a decent universal platform. I'm still not sure if this buys anything over the $12 Allegro microboards for these desktop devices.

After looking at the amperages involved I am a little concerned that the 40-pin parallel IDE port won't work so well, though.

[r691175002]

It is generally worth paying more for stepper drivers if you want high speeds. Reasonance/inductance starts becomming an issue as you push the system to its limit and newer drivers will compensate.

A machine is limited by the weakest link so every aspect you mentioned is improved. There will be linear rails, highly preloaded belts and servos. Mill accuracy is a fairly large topic but literally every aspect from the shape of the castings to the exact type and number of bearings used in every location has an impact.

Most hybrid steppers are brushless servos with a high pole count (or a stepper motor being driven by a servo driver). Stepper motors and brushless motors are essentially identical and the differentiation is purely based on how you use them.

For what its worth, AFAIK DMM tech servos are price competetive with hybrid stepper systems and have much better performance. Not that I would recommend spending that kind of money on the type of machine being discussed.

[macona]

About the only thing similar about stepper motors and brushless motors is that they are both are electric motors. The construction and characteristics are significantly different.

As for unified machines, one machine that does everything does everything poorly.