Performance Benefits of Cloud Control?



Last night I was pondering dan’s statements about how having control processing done through Cloud computing and not relying on Gcode allows a lot of cool capabilities. After a quick study of Gcode it’s clear that it is very limiting. I wondered whether the typical laser cutter control board does a significant amount of intelligent interpretation of the GCode commands. What I don’t know, and hope the community helps with my education, is whether a typical laser cutter (Not Glowforge) uses intelligent control. I’ll give you the most simple example I can think of. If you command the laser to cut at a basic right angle (move the head some distance X and then an equal amount Y at some power/speed) does the control board simply execute the command or does it interpret the command for best effect. Does it cut for some distance at one speed, adjust for kinetic effects of the head as it approaches the corner, adjust the power to compensate for the speed difference to avoid burn out of the corner point and then change direction adjusting power and speed for best effect? If not the GF approach seems to open up many possibilities. I can probably think of dozens of more complex use cases.


That greatly depends on the controller. The stock K40 laser does not do any motion planning. But some 3rd party controllers (i.e. GRBL, TinyG(and G2) & LinuxCNC/MachineKit does. The other benefit to the latter is that you can configure the accel/decel of the gantry to reduce that snapping action when executing commands.

The main advantage to the GF model is that they can control/update/improve all of this in one shot. v.s. the hours it usually takes to do this (at the enduser level) i.e when I switched from TinyG2 to LinuxCNC it was about a day’s worth of trial and error to get exact distances correct and configuring the accel/decel values.

So it’s like having a MESA class motion controller (Which is a dedicated controller that runs on a FPGA which costs a lot of money) available to all the Glowforge’s without having to include that cost into the unit.
With that type of saving- then can use that on better optics and the magical lid camera and the macro cameras.


LinuxCNC and Mach3/4 both have trajectory planners which handle this sort of thing and I believe that their effects can be tuned by the user in the configuration files.

I’m not sure, but believe that both also have the capability to be customized with macros or other means to control devices or to provide special functions, such as control a cooling or ventilation system or other laser cutter accessories.


Never really thought about it because all of my laser uses have been fairly simple and forgiving. Never needed to investigate what the controller was doing. The benefits related to firmware and Cloud based S/W updates, configuration control, and even the economics of future fee based capabilities are clear. I also have a my own vision (right or wrong ) of how the motion planning and other control functions might be done by Glowforge. What I was missing, and you answered, is how capable are other laser cutter control boards. Thanks.


I come from the woodworking cnc world. I understand with having a spinning blade driving into a solid with a few hundred or thousand pounds of gantry behind it needs to slow for corners and such. With such a small gantry and no lateral forces on it from cutting, how much control on speed and feed in corners and such is a concern?


Regardless what tool is at the tip of the Z axis (or centerline of the X/Y axis) or the size of the gantry. It all has mass. And when moving generates inertia. The better you can control that. The better the resolution.


Stepper motors aren’t quite what you are used to. All by themselves they can have characteristics that require acceleration ramps up and down. Overshoot, resonance and other design issues are a nightmare. Designed a lot of low voltage circuitry, but have the sense to stay away from motor control. The precision requirements of the GF is also fairly small. Doesn’t take much.