Would This Require Custom Kinematics?
-
Hello, I am brainstorming layouts for a large H/CoreXY printer, and would like to mount bowden extruders such that they travel in the long direction.
Motor 1 would be expected to follow the position of the gantry. Motor 1 is likely to be a servo that has been acceleration/torque limited, so it will desync slightly from the gantry but never stray too far.
Would this be easy to configure?
-
I'm not sure that the existing firmware can do that, but the new generalised Cartesian kinematics in my current firmware build certainly can. You may need to switch the kinematics within the homing files so that you can home the extruder "axis" separately.
-
Thanks, do you have any documentation on the generalized cartesian kinematics or is it still too new?
Also, would it be possible to home the extruder and corresponding CoreXY axis simultaneously? The extruder carriage must stay within a few inches of the gantry or it will pull on the bowden connection.
(In practice there will be a mechanical stop so a desync will fault the servos.)
-
I don't have any documentation yet. Basically, you use the M669 command to define a matrix of coefficients saying how much each motor needs to move for a unit movement of each axis. By changing the coefficients it can support Cartesian, CoreXYU, CoreXYUV, CoreXZ, MarkForged, variants of these with additional axes, and any other kinematics for which the movement of each motor is a linear combination of the movement of the axes.
You would be able to do the initial homing of the extruder axis and X simultaneously; but if the X endstop is triggered first then the extruder axis motion would stop too. So you would need to use a combined homing move followed by an extruder homing move.
-
Thanks dc42, one last question: How hard would it be to omit a motors acceleration and velocity limits from the trajectory planner, while still obeying the limits in the generated pulse train?
Just in case that question doesn't make sense:
If I define custom linear kinematics, both the extruders and hotend will move in lockstep. But I don't actually want them to stay aligned - the extruders are much heavier and the bowden tubes will accommodate misalignment. It is better if the extruders lag behind the hotend.
However if I configure the extruder axis to a lower acceleration within duet, the whole system will obey the that limit (I think?). So to properly desynchronize the axes, the extruder axis pulse train must be smoothed outside of the Duet, probably using a programmable servo drive - expensive and complicated.
If possible, I would like to omit the programmable servo drive and configure a lower acceleration limit within Duet without handicapping the overall system. Would it be possible to add some kind of flag which "this axis does not need to constrain overall motion" flag to the firmware, or would it be too convoluted?
This would also have applications for printers such as: https://somei3deas.wordpress.com/my-corexy-printer-build/5-colour-upgrade/
-
@691175002, that would be a very major change, not possible in the standard firmware build due to the extra RAM it would need. Possible in a special firmware build with e.g. reduced maximum numbers of axes and extruders to reduce the RAM needed.
I suggest you do the mechanical design so that you can use the same acceleration for the extruder axis. It won't matter if you get some ringing on the extruder axis, as long as it isn't transmitted to the XY axes.
-
@691175002 As the designer and builder of the printer you linked to, may I ask what is it exactly that you hope to achieve? If you read some more of my blog, you'll see that I have demonstrated real world printing at 300mm/sec. That's with a total moving mass in excess of 3 Kgs. The kinematics of having the two gantries linked are not a constraining factor.
Before I upgraded from 3 extruders to 5, the upper gantry was passively driven by the lower gantry, with a certain amount of decoupling. This did nothing to improve print speed or print quality but the constant flexing of the Bowden tubes and wiring caused all sorts of problems.
So trust me - I've been there and done that and having the upper gantry exactly follow the lower gantry is far better.
-
@deckingman said in Would This Require Custom Kinematics?:
@691175002 As the designer and builder of the printer you linked to, may I ask what is it exactly that you hope to achieve?
I'm hoping to build a multimaterial printer (~4 filaments) with a wide print area. Something along the lines of 30"x10". This is probably too large for a static bowden system, but if the extruders follow in the long axis the bowden tubes can be very short.
I'd like it if the rectangular build area let me get away with moving the extruders on only a single axis, since that is much simpler than duplicating the full XY gantry.
This project is still in the early stages, but given that the axes are mechanically separate it seemed worth investigating the possibility of using different accelerations. It might make the printer quieter, if nothing else.
-
@691175002 I might be worth investigating simply having the extruder gantry being passively "dragged around" by the main hot end gantry, especially if the Bowden tubes will be as you described. I tried this and it was surprisingly successful. One advantage is that for smaller zig-zag type moves, the extruder gantry actually remains static as the Bowden tubes absorb the movement. It was only when I upgraded to 5 extruders that the engineer in me started to weep when it became apparent that I would be using a smaller mass to drag around a larger mass.
The other thing I tried was having the extruders suspended above the centre of the bed on a sort of pulley/counter-balance arrangement. A bit like a flying extruder on Delta. That was less successful.
If the extruder gantry is going to be actively driven, then you need to consider how to home it. If you don't home it, then there is a real chance that it will get knocked out of position when the printer is powered off. Then when you turn the printer on and attempt to home the printer, the extruder gantry hits the frame before the hot end gantry hits the end stop (been there, done that, got the Tee shirt).
Running different accelerations would also run the risk of the gantries getting out sync and leading the to same relative positional problems.
-
Hi,
Have you considered using a remote drive extruder like the Zesty Nimble?
Frederick
-
@fcwilt said in Would This Require Custom Kinematics?:
Hi,
Have you considered using a remote drive extruder like the Zesty Nimble?
Frederick
That's another good option and one that I would have given serious consideration to if they had been available when I built my printer. Of course the OP would need 4 but the cost would be offset against the need for a second gantry.
