views on "CoreXY" ?

Discussion in 'BigBox General Chat' started by R Design, Aug 14, 2016.

  1. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    In a recent thread (https://forum.e3d-online.com/index.php?threads/well-no-big-box-for-me-then.1900/) someone brought up the CoreXY principle ( http://corexy.com/theory.html ).

    Of course little me hadn't heard of it but I think it's super interesting.

    Despite spent many hours looking at the clever ultimaker design which allows the moving parts to be very light, it never occurred that there might be a solution involving belts.

    Does anyone have any experience of this?

    What are the disadvantages of CoreXY?
     
  2. GeckoBox3D

    GeckoBox3D Well-Known Member

    Joined:
    Aug 5, 2016
    Messages:
    55
    Likes Received:
    11
    I'm currently designing a CoreXY machine actually. The benefits are huge. Now that we have the E3D titan you can have a 200g or less direct extruder on the carriage, making huge weight savings on the moving gantries without the downsides of Bowden. It's easier to calibrate than delta, and alot more rigid, making results highly repeatable.

    The printer size to build size ratio is better too than any cartesian (perhaps even delta?), making a true desktop machine in my eyes. Easy to enclose aswell,so you can actually do decent ABS prints in a heated chamber.

    The only issues I can see, is the large belt runs. Even on a modest build size, the belt runs can be meters long. This is easily offset by using 9mm wide gt2 belts,or steel core belts on large pulleys (a necessity if using steel core, but can lower resolution).

    Secondly, the belts will expand more in a heated chamber. This is only an issue if you want 60c+ chamber and if you do than belts are the least of your problem. Properly specced gt2 belt has an optimum operating temperature up to 80c, but than so do most linear movement applications like rails and rods. Moving above 80c requires some serious cash outlay for decent linear rail which won't lock up with the thermal expansion. But than, this is PEEK And ultem territory. Here be dragons as E3D says.

    There are also issues with properly and evenly tensioning the 2 belts. This only ever has to be done once though (or once in a blue moon)
     
  3. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    All things considered, it is taking a long time to catch on then?

    So you just add a couple of lines to existing firmware so the motors behave correctly?
     
  4. moshen

    moshen Well-Known Member

    Joined:
    Dec 20, 2015
    Messages:
    90
    Likes Received:
    12
    I'd love to see a CoreXY adaption for the BigBox using the same motors rails and mounting holes. I think it's possible!
     
  5. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    Yeh I was pondering that it would only take minimal redesigning/reprinting of the X/Y carriage parts to convert the BB. But at the moment I'm still trying to decide what is the lowest hanging fruit in terms of pushing the print quality/speed ratio.
     
  6. mike01hu

    mike01hu Well-Known Member

    Joined:
    Apr 22, 2014
    Messages:
    957
    Likes Received:
    166
    The principle is really good and further mass and drag savings could be achieved by using CAN bus for the electronics to minimise the cable run size, but a new approach to the controller and a super fast processor is essential; this is the province of better people than me.
     
  7. GeckoBox3D

    GeckoBox3D Well-Known Member

    Joined:
    Aug 5, 2016
    Messages:
    55
    Likes Received:
    11
    I'm not sure why it hasn't picked up, but marlin has core xy functionality already I believe. Reprap firmware (duet boards) certainly does, as does smoothieware.

    It's more difficult to implement than cartesian certainly, but it's not a huge hurdle. The design I'm doing at the moment uses v slot, and uses metal for nearly all components. No acrylic panels needed for rigidity, or plastic corners. Placing the belt so that it could run straight and not be in the way was awkward but hardly much of a hurdle. I'll post pics of the design soon.
     
  8. Nuno João Casteleira

    Joined:
    Aug 13, 2016
    Messages:
    2
    Likes Received:
    0
  9. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    I thought about a bus, but CAN is probably not a great choice as it's patent encumbered so not so compatible with FOSS projects like Marlin. There are specific licence restrictions on burning it to things like ASICs so a lightweight dedicated controller is a no-no, you'd have to implement it on a micro at each point :(

    It's also a bit overkill I suspect, I'd go for something simple. We don't need all the message handling and QoS/priority capability in CAN because we don't need to interact with that many devices (unlike all the targets in a modern car)

    Also really it's the flying weight up on the X-Y axis where you want to save cabling and by the time you have taken up enough power to run the hot end and the extruder, moved the heater FET's, moved the stepper driver next to the motor and added a bus controller to work that plus the temp sensors, plus levelling, plus fans. half of your RUMBA board is going to be flying around at the top :) ..I'm sure the moving parts would get heavier not lighter :)
     
    mike01hu likes this.
  10. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    Of course, crazy crazy would be to use the two linear rods as a power bus, overlay digital signals on the DC offset to run the controllers, move bits of controllers up there and if you could solve the problem of needing good electrical contact on something that is also lubricated then you'd have a cable-less print head :D
     
  11. mike01hu

    mike01hu Well-Known Member

    Joined:
    Apr 22, 2014
    Messages:
    957
    Likes Received:
    166
    Just throwing thoughts into the conversation:rolleyes::)
     
  12. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    I'm surprised nobody has had a proper go at standardising a stepper motor with an integrated controller yet (with some sort of addressable bus like i2c)

    Stepper wiring looms are a total pain as it's lots of high energy switching so in terms of electrical noise is a bit of a pig if you are trying to achieve any sort of EMR spec for CE/FCC or MDD compliance, plus we'd gain the capability of addressing up to 127 steppers without running out of controller pins.

    Maybe I shouldn't say that before trying to get a patent on it first :)
     
    mike01hu likes this.
  13. R Design

    R Design Well-Known Member

    Joined:
    Oct 7, 2015
    Messages:
    918
    Likes Received:
    183
    I would have thought that the only stepper for which we really care about lightening the cabling is the extruder motor in a non-bowden setup?

    But a stepper with an i2C controller will also have 4 wires, like a standard stepper, (Power, GND, SDA, SCL) So no reduction in load to the hotend only the cable run down from the motion system to the controller board (since a shared i2c circuit for all). Unless we cabled the fans to the same microprocessor...?

    I reckon a good way to simplify wiring would be to mount the controller board on the back or the side of the machine at the height of the motion system since that's where 90% of the wiring (nozzle, fans...) leads. It would also make it more accessible (and lead to placing the LCD elsewhere). On the BB I'd convert the space defined by the Z motors depth into a well for storing hex keys, nozzles, calipers, gluestick and the like.

    Last week I saw a guy running a i2c line closeby a linear actuator. He had terrible interference problems from the motor which he was attempting to solve by using a MAX232 (RS232) chips so as to be able to send the signals at a higher voltage. Don't know if that was the smart thing to do?
     
    mike01hu likes this.

Share This Page