While most RepRap and 3D printer controller boards use an ATMega or other 8-bit microcontroller, the Smoothie uses a 32-bit ARM chip in the form of an NXP LPC Cortex-M3 chip. Not only does this allow the Smoothie to do some very cool things with your machine – native arcs and circles, for example, but this better hardware also allows for Ethernet, drag-and-drop firmware, and exposing the USB port as both a serial port or mass storage device.
http://hackaday.com/2013/09/30/smoothieboard-the-be-all-end-all-cnc-controller/
Native arcs and circles?
So we finally need slicer software that reads CSG CAD files because STL canot describe arcs so the proper g-code for arcs cannot ever be written by the slicer.
I’m not sure I like to have the stepper drivers integrated into the board .I’m currently replacing a burned out stepper driver.
Also that means that the stepper voltage is fixed and you cannot choose between cheap 12V, 36V or hefty 48V.
The usual approach with polygonal models is to first generate a tool-location point based toolpath - consisting of many many short line-segments, and then run a filter that detects, within tolerance, your curves of choice (arc, spline, nurbs, etc) from the point data.
Not very elegant or effective, but writing robust CAM functions for general CSG shapes is really difficult.
@Anders_Wallin The “polygon models” part is wrong.
FreeCAD is working on the robust CAM part.
I’m trying to help by integrating my existing 4+5 axis code.
CAM strategies for 3D printers are trivial.
You never collide with anything because it doesn’t exist yet.
https://sourceforge.net/apps/phpbb/free-cad/viewforum.php?f=15
Wow, the number of people hating on this on hackaday is amazing.
be-all, end-all, cnc-controller is a bit grand.
for a specific purpose (stepper motors, none or very small cutting forces, pwm heaters required, etc, etc) it might be OK.
@Matt_Kraemer1 ugh, removing smt parts makes Eric angry! Apparently traces are don’t like to stay on boards, so Eric sets them free… To the scrap pile.
But in all seriousness, mounting the motor controller directly to the main board is a much better heat sink, although at a high risk as mentioned.
How are you guys frying your steppers? I have a Rambo and hooked it up properly but I want to keep it from frying !
I can’t see any support for the LCD & click encoder; a feature I’ve enjoyed using for making adjustments and selecting files to print without a computer connected.
I can see this board having an advantage on a multi-purpose machine. A bit expensive for a single purpose machine.
Actually, it will support LCD panels with click encoders and there is a special Smoothie Panel with extended capabilities. See the site for details, http://smoothieware.org
I’ve been using my Rambo on 24V setup, heavily for several months now with no hiccups.
How are you people frying them?
But as above, the smoothie can have all the processing power it likes but till there’s a new slicer and file format it does no real good.
The M$ are working on one apparently, not sure on my feelings about that.
From what I understand, if you move your stages by hand to fast, it cause the generated current to flow through the controllers, occasionally cooking them
good to know, thanks!!
also, don’t disconnect motors while the thing is powered on.
I’ve had an integrated board for over a year and had no problems, I’ve seem people fry chips left and right with RAMPs b/c they don’t check polarity or don’t seat it well enough before they turn it on. And the cost integrated is still about the cost of a RAMPS from a ‘reputable’ resource.
I still tend to agree with @Marcus_Wolschon , for CNC machining, you’d want the flexibility of beefier motor controllers. That being said, I’d be nervous to run larger motor current through PCB traces through the stepper ic. I think I’d be more comfortable sending the high current through a MOSFET, similar to a Darlington transistor, I believe it’s referred to as. Pretty much drive the fet with the allegro controller output. Then you could have your high power rail and your motor output connected directly to the MOSFET with terminal blocks. I guess you’d have to make sure the MOSFET could respond fast enough to keep up with the pwm. It’d be cool if the terminal blocks and MOSFET were attached to the motor controller, so it could be a drop in replacement. Not sure if there’s room though as now you’d need to cool both the MOSFET and the allegro.
Do i smell a Kickstarter 
@Eric_Moy it’s not quite as easy as that - stepper drivers also need a current feedback to properly work. So if anything, we could design a board that plugs into the Pololu sockets and routs step, dir, enable and gnd to an external driver, but really that wouldn’t be much more than a bunch of wires.
There are breakouts on the PCB for each stepper that could be used to control an off board driver. You might even be able to use that to bypass a fried driver chip but it might also require cutting some of the traces on the board to isolate the bad chip.
@Matt_Kraemer1 even though Allegro- or TI-based drivers might not have the kind of smartness of a Gecko driver, they still employ basic current regulation (á la buck converter). This is what the two sense resistors are needed for.
@Thomas_Sanladerer couldn’t the current sense resistor be measuring the MOSFET current. As long as the resistor was sized correctly, you could trick the speedo into thinking it’s its own output. And I was thinking all this would be on the pololu, and the terminal block would just be 2 power in leads, and 4 motor out leads. Doable?