Flying extruder with 4th axis

sungod3k

Interesting upgrade. Since discovering deckingmans machine I have been thinking about that issue.

I have a 1m tall kossel style printer with smart effector and the bondtec QR (the one with geared motor), so I first did a classic flying extruder with rubber bands connected to the carriages, but there was too much bouncing around.

So I started using counterweights (pic on thingiverse). On top of each tower I have now one of these https://www.thingiverse.com/thing:3467741/ that use two grove bearings to connect a little bucket filled with plaster of paris that rides up and down the towers. That balances the extruder quite nicely.

I wonder would you would say is the advantage of motorized version of this is?

Cheers

dc42

@jooil said in Flying extruder with 4th axis:

@dc42
I understood.
I just want to know if you have some plan to add offset function to the new axis.

I hadn't planned to , but it could be done. If you used it then when you home the printer, the distance between the extruder and the effector would be incorrect (by the amount of correction you used) when you home the printer; but it would be corrected when you perform the first move (e.g. the 5mm downward motion that is typically at the end of the homedelta.g file). Is that really what you want?

deckingman

@sungod3k I used to use the "pulley and counter-weight" system on my CoreXY, with the extruders suspended above the hot end, from a pulley that was fixed high up, roughly in the centre of the bed. It was "kind of OK" in that it allowed me to use shorter Bowden tubes but when you think about it, every XY move has to move the mass of the counter-weight vertically. Then I changed to a separate XY gantry to hold the extruders but this was passively "dragged around" by the hot end gantry. This was an improvement but that's a lot of mass to be moving around because then you have the mass of the rails and carriages as well as the extruders themselves, and accelerations were limited. When I upgraded to 5 extruders, I added even more mass but more importantly, I was effectively using a small mass to drag around a larger mass with some degree of de-coupling between the two. The engineer in me just couldn't accept that (the inertia of a large mass will have a severe impact on the small mass coupled to it when a change of direction occurs). Also, because of the de-coupling between the upper and lower gantries there was a fair amount of flexing of the wire to wire connectors for the temperature sensor and heater which lead to a few failures.

So having the extruder gantry motorised, effectively means that I have my 3kg total mass divided between 2 gantries, each which their own 2 motors. Which means I can run twice the acceleration. Also, with them being perfectly synchronised, there is no flexing of the cables that run between the two so I have no more joint failures. Having said, the small amount of decoupling that I had when the extruders were passively driven did mean that for small moves such as zig-zag infill, only the hot end moved and the extruders effectively remained stationary.

So for me on my CoreXY (actually CoreXYUV now) having an actively driven extruder gantry is a better solution than a passive one. I would imagine that would be the case for Deltas too but I know next to nothing about them. If the existing motors will accelerate the print head as fast as you want, while still passively dragging extruders around, and there is no danger of the mechanism causing the effector to tilt, then there is probably no point in motorising the extruder gantry/tower. If on the other hand, the coupling to the extruder carriage could cause the effector tilt, or acceleration is being limited by the motor/mass combination, then using a motor to power the extruder gantry/tower will be beneficial.

Jooil

@dc42 I Not really want the offset function anymore.

I added the following G-code to homedelta.g as your advice.
G1 Z-5 F6000 ; 5mm downward motion.
G1 U-15 F1800 S2 ; fine tune U_axis' carriage position. move down 15mm

Now I can finetune U_axis' carriage position.
Thank you very much.

dc42

@jooil said in Flying extruder with 4th axis:

@dc42 I Not really want the offset function anymore.

I added the following G-code to homedelta.g as your advice.
G1 Z-5 F6000 ; 5mm downward motion.
G1 U-15 F1800 S2 ; fine tune U_axis' carriage position. move down 15mm

Now I can finetune U_axis' carriage position.
Thank you very much.

I think you may find that the U carriage reverts to its standard position when you next enter a regular move.

Jooil

@dc42 Other towers do act as you mentioned.
But U_axis doesn't seem to revert to the standard position.
It acts like -15mm offset. If I adopt a magnetic end_stop to avoid a collision, I could even set a positive offset I suppose.

dc42

I'll extend M666 to allow UVW endstop offsets in the next beta release.

Could you make a video of your machine printing, that I can share with others?

Jooil

@dc42
I will in the future.
Currently, I am not prepared for the heat bed. So actual printing is not ready to show. And the 4th tower is a bit ugly.
I can take a video of bed leveling(G32,G29) though.
Is there some dryrun G_code in Reprap firmware? I can't find it.

dc42

I do dry runs simply by not putting any filament in the extruder.

sungod3k

@deckingman said in Flying extruder with 4th axis:

So for me on my CoreXY (actually CoreXYUV now) having an actively driven extruder gantry is a better solution than a passive one. I would imagine that would be the case for Deltas too but I know next to nothing about them. If the existing motors will accelerate the print head as fast as you want, while still passively dragging extruders around, and there is no danger of the mechanism causing the effector to tilt, then there is probably no point in motorising the extruder gantry/tower. If on the other hand, the coupling to the extruder carriage could cause the effector tilt, or acceleration is being limited by the motor/mass combination, then using a motor to power the extruder gantry/tower will be beneficial.

Right, I hadn't considered the XY movement. Having the motor quietly resting above the effector makes sense.

I had considered building a scaffold on top of the delta, so the extruder gets strung up by only one tether, but that would maybe lead to more wobbling since 3 tethers with their friction can calm down XY motion better than just one.
As it is, I dont see an issue. Since my frame is weak im accelerating not that hard, but it could become in issue adding more inertia to moves at higher speeds.

As a third option i also had thought about fixing the extruder to the carriages, but that would have meant not decoupling the extruder weight from the carriages.

Would be interesting to see if one can add a rod on top of the printer parallel to the towers that is still counter balanced but doesnt allow the extruder to move in XY.

In any case: a good opportunity for @ander to put together a upgrade kit

Jooil

@dc42 I made some video of 4th axis flying extruder.
Because I am not used to video editing, I uploaded 4 separate videos on youtube. sorry about that. Pls, visit followings.

4th axis flying extruder overview... https://youtu.be/Ig0k9NiVD94
Side view ... https://youtu.be/hCjmkRbwGWc
Top view ... https://youtu.be/Z_L8L-FmCf0
Carriage view ... https://youtu.be/bgNJpSbPvUQ

If you need other scenes, just ask.

dc42

Your last 2 links are the same. But the second link shows the 4th axis very well. May I post it on other sites?

Jooil

@dc42, Of course, you can post it on other sites.

The PTFE tube length is 280mm from extruder fitting to effector fitting.
This is significant improvement comparing to my mini KOSSEL's 450mm.

I corrected the link of the carriage view, the 4th video.
I attached a metric ruler on the 4th rail. So you can see how far and fast the extrude carriage moves up and down when effector is traveling the same z_height XY plane.

My current setup is as follows
U_axis rod Length is 300, U_tower position M669 X0 Y-100.

The carriage moves dully relative to XY position if the U_axis rod length
is set high(ex. X665 L358:358:358:600), whereas the movement is very rapid and far when U_axis rod length is set low (ex. X665 L358:358:358:300).
If I set the rod length too short (for example 200mm), the effector cannot reach U_axis' opposite(diagonal) edge. And the carriage moves too fast enough to skip steps. Firmware restricts the movement.

The ideal U_tower position seems to be the center[M669 X0 Y0] given that fitting do not wear out and the PTFE tube don't flex too much. I will try this when I am ready for gimbals similar to ZATSIT's FlyStruder.

ander

hi there!!!
Sorry i was kind of busy with work and disconected from all this proyect, i still didn´t make any video(i will) but it is working fine so far...but i have a little problem using slic3r prusa edition. If i generate the gcode with simplify3D there is no problem, but with the prusa edition it goes so slow that it looks it is not even moving, i mean all the movements, homing, auto level, printing...i tried to higher all the speed settings that the slicer brings but the result is the same. the simplify version i´m using is 2 years old and the prusa slic3r is the latest. i was wondering if the 4th axis´s speed is limiting here because in the slicer i have no option to change it

dc42

I don't think it's related to the 4th axis. I suggest you check whether the slicer has inserted any M203 commands in the GCode file that you are trying to print.

Did you choose the correct GCode flavour in the slicer?

ander

Hi there, i don´t really know what you mean by flavour, the prusa slicer just has a button that gives you the g file and that´s it. i kind a feel like i´ve tried every setting now, dunno what happens

deckingman

@ander said in Flying extruder with 4th axis:

Hi there, i don´t really know what you mean by flavour, the prusa slicer just has a button that gives you the g file and that´s it. i kind a feel like i´ve tried every setting now, dunno what happens

On PE Slic3R go to the "Printer Settings" tab, then click "General" which you will find at the top of the list on the left hand side. The on the right hand pane you'll find a section entitled "Firmware" and under that is "G-code flavour" with a drop down box. Select "RepRap/Sprinter" from the list. Then save the change.

Phaedrux

Which version of slicer PE specifically?

In the print settings tab under speed, there is an acceleration control. I believe it's hidden now by default in the latest 1.42 beta and only visible when set to advanced/expert view.

Perhaps you can post the gcode file that has this problem?

ander

Hi there!!
you guys were right, i´m used to program in Marlin for arduino, and i put marlin instead of reprap, thanks for helping. i couldn´t post earlier because for some reason the forum was blocking my messages again, it looks that works fine now.

ander

Hi there!!!
a bit late but here goes my setup:
https://www.youtube.com/watch?v=ToBLN7YHWt4