Weird Laser Kinematic Ideas - Firmware suitability?



  • Hi DC,

    Now that I've got such promising news about how far duet has come with laser support, its time to think kinematics for the machine. I don't need a motorized Z axis, which makes things simple, but it needs to have a large cutting area, and otherwise be as small as possible because I live in a city apartment; certainly it needs to have a build volume larger than its own size.

    My two ideas are: a typical x gantry on which the laser slides left and right, with the ends of said axis attached to two "carts" with wheels controlled by steppers, essentially just driving back and forth for the y axis. That would be a typical cartesian for firmware purposes, with steps/mm being set based on the mechanism between the motor and wheels and the wheels size, so not worried about that idea from a software perspective.

    Ideally, I would have a cart on coasters that somehow just drives itself on both axes and at any angle in between. Possibly this could be done with a cart riding on coasters/ball bearings with one stepper driven wheel which in turn is mounted on a second stepper that rotates the driven wheel in the direction in needs to go. This wouldn't be conducive to printers so I don't think any such firmware exists. Thought I would pitch the idea to DC42 and see if one of the current configs could be modified to handle such an arrangement where one motor aims and another drives. Let me know. Omniwheels could work in a standard cartesian setup, but I'm not picturing that approach having much in the way of repeatibility or accuracy.

    The more feasible variation is basically a modified hang printer. The cart holding the laser would be on ball bearings that can move in any direction as above, but it would have motorized cables that pull the assembly around. Because it doesn't need a Z, it'd be a good deal simpler than a hang printer. My question with this approach is how easily could I configure duet to handle this arrangement. I should be able to just disable the stepper that would handle the z axis, throw a G92 Z0 at the bottom and config and be good, so long as x and y motion don't rely on the z stepper for x/y positioning. My hunch is that they do not, and that they pull themselves in a delta configuration around x/y and the z motor exists to feed out enough line to allow the assembly to reach the right x/y location without swinging it up like a pendulum. Any thoughts on this approach?


  • administrators

    @clearlynotstef said in Weird Laser Kinematic Ideas - Firmware suitability?:

    Hi DC,

    Now that I've got such promising news about how far duet has come with laser support, its time to think kinematics for the machine. I don't need a motorized Z axis, which makes things simple, but it needs to have a large cutting area, and otherwise be as small as possible because I live in a city apartment; certainly it needs to have a build volume larger than its own size.

    SCARA printers are good at that, but they have other problems.

    My two ideas are: a typical x gantry on which the laser slides left and right, with the ends of said axis attached to two "carts" with wheels controlled by steppers, essentially just driving back and forth for the y axis. That would be a typical cartesian for firmware purposes, with steps/mm being set based on the mechanism between the motor and wheels and the wheels size, so not worried about that idea from a software perspective.

    I don't understand what you mean by two carts, please explain.

    The more feasible variation is basically a modified hang printer. The cart holding the laser would be on ball bearings that can move in any direction as above, but it would have motorized cables that pull the assembly around. Because it doesn't need a Z, it'd be a good deal simpler than a hang printer. My question with this approach is how easily could I configure duet to handle this arrangement. I should be able to just disable the stepper that would handle the z axis, throw a G92 Z0 at the bottom and config and be good, so long as x and y motion don't rely on the z stepper for x/y positioning. My hunch is that they do not, and that they pull themselves in a delta configuration around x/y and the z motor exists to feed out enough line to allow the assembly to reach the right x/y location without swinging it up like a pendulum. Any thoughts on this approach?

    Again, I don't follow exactly what kinematics you are proposing.

    Note, any kinematics for which the motion of each axis is a linear combination of the motion of individual motors is now supported in the standard RRF 2.03 and 3.x binaries.



  • I would think that a CoreXY arrangement would be the ideal for a plotter style system like a laser cutter. The frame can be made very rigid very cheaply, the effector can be kept very light and travel very quickly.

    But maybe I'm missing something.



  • @phaedrux the whole objective is to print as large as possible with as small a device as possible. For as often as I'll be using it, and as large as it would need to be, a corexy for this project just isn't feasible in a nyc apartment.



  • @dc42

    The first idea would look something like this (obviously rough mock-up)alt text

    i.e. typical x gantry mounted on a y axis that literally just drives th whole thing back and forth

    The second idea, which is an even rougher mockup (i wouldn't use cart casters but it gets the point across) is just what it sounds like, a hang printer on wheels:
    alt text


  • administrators

    The first one would use standard Cartesian kinematics, The second one would need kinematics written for it, essentially a 2D version of Hangprinter.

    Both idea worry me rather. I can imagine the card rolling off the table and the laser burning something that it shouldn't (possibly you).



  • @dc42
    Funny how people think about things differently. One of my main positives of the second one is safety, it's small enough that it could be completely surrounded with a skirt like a brush or vinyl that goes all the way to the medium and no light would even escape. Obviously if used on a table it should have bumpers on all sides, though it's more likely I use it on the floor, personally.

    The real question boils down to how do hang printers work? If the z was controlled (by means of wheels in this case), would only using the 3 xy motors and leaving the z unplugged still result in proper xy movement (if using G92Z0 first so the machine considers itself homed), or does xy positioning actually rely on the z motor for more than moving the height and feeding enough line to make the xy moves reachable.



  • @clearlynotstef said in Weird Laser Kinematic Ideas - Firmware suitability?:

    ...it's small enough that it could be completely surrounded with a skirt like a brush or vinyl that goes all the way to the medium and no light would even escape.

    When working with metal laser processing I've seen smoke & particals have a detrimental effect on laser quality.

    How are you planning to control this and keep your optics clean if the work piece is on the floor and there is a skirt all around the cart?



  • @doctrucker good question! For this particular project, the medium is actually always underneath glass, specifically to keep it from burning as much as possible by limiting oxygen access, and also to keep it held down and flat. So the cart contraption would be riding on top of the glass and cutting through it.



  • What kind of wavelength and power level are you intending to use?

    I tried engraving through glass with a 2.5w 405nm laser, and I'm not sure if it was the heat from the work piece or the laser directly interacting with the glass but the glass got damaged along the lines. My intention was to help keep the work piece flat, but had to abandon that idea.

    (May revisit it for trying to cure UV soldermask, but that would be at much lower power levels, like 20mW or so)


  • administrators

    Bear in mind that the glass will reflect a significant fraction of the laser light back towards the laser, which could damage it. I presume you are using a visible or near infrared diode laser. CO2 laser light won't penetrate glass.



  • Hm, thought the UV spectrum around 400nm would pass though glass, but seems not to be the case, but that was the main reason to ask OP for the intended wavelength.


Log in to reply