New REXY Kinematics on duet firmware possible?

zabana

Hi!
I have found this new kinematics to drive a direct drive extruder but with extruder motor outside of the axis like COREXY ones. In fact the printer works in corexy mode.
The extruder movement is dependant on corexy ones so my question is, is this possible right now with actual firmware?
Here is a link to the youtube video so all of you can see it how it works and how clean and powerfull system is in my opinon.

Video

dc42

RRF doesn't support that type of extruder yet. I will need to think about how to implement pressure advance for it. In principle you could have a combined XYE movement that resulted in extrusion without any E motor steps, until you apply pressure advance. So the pressure advance can't just be applied to the E movement as is usually the case.

zabana

Well, maybe is a good idea start without pressure advance due this extruder is direct drive and probably will not need it or a minimum one just in case.
Also I thought it will be easier because the duet firm needs to calculate the E movement with the pressure advance and after add or substract movement depending on the corexy axis movement.

oliof

I'd be interested to hear more about this extruder-inclusive kinematics. Any place to read up on it?

fcwilt

Interesting idea.

I don't fully understand why the Z axis needs such a relatively complicated belt layout.

Frederick

o_lampe

@fcwilt said in New REXY Kinematics on duet firmware possible?:

I don't fully understand why the Z axis needs such a relatively complicated belt layout.

It looks like a winch to double up lifting-power?

o_lampe

@zabana said in New REXY Kinematics on duet firmware possible?:

direct drive extruder but with extruder motor outside of the axis

Just wondering if the long belt is better than a long Bowden tube. Definitely better than the remote cable driven extruders IMHO.
I'd like to see how they translated the vertical rotation of the filament-drive wheel into the usual horizontal filament gear orientation. Please tell me, they don't use a worm gear

rabbit

@o_lampe they use a bevelgear

fcwilt

@o_lampe said in New REXY Kinematics on duet firmware possible?:

Definitely better than the remote cable driven extruders IMHO.

I've got three printers using Nimble v1.2 units and they work just fine.

The v2.x units did not work well at all.

Frederick

fcwilt

@o_lampe said in New REXY Kinematics on duet firmware possible?:

@fcwilt said in New REXY Kinematics on duet firmware possible?:

I don't fully understand why the Z axis needs such a relatively complicated belt layout.

It looks like a winch to double up lifting-power?

That would explain the belt layout - but why not just use a stepper with the needed gear ratio?

I'm not a fan of long belt runs.

Frederick

jimmyoh

@dc42

I think the idea is to get away from a long bowden tube and eliminate the need for pressure advance?
At least this is what i'm trying to accomplish.

Jim

Please keep up the good work, I use the Duet3D on all my Printers and a few other things ;-D

amythebun

@zabana hey did you try this idea? I was debating doing the same on my new printer. I'm worried the long belt will cause some unforeseen issues, I personally would either do CoreXE or a markforged IDEX style one (the second X would be continuous and move the extruder)

also I don't get why RRF doesn't support these mechanics, wouldn't they just be a regular cartesian? with the expection of X moving 45 degrees if the y stepper isn't locked.

T3P3Tony

@amythebun said in New REXY Kinematics on duet firmware possible?:

also I don't get why RRF doesn't support these mechanics, wouldn't they just be a regular cartesian? with the expection of X moving 45 degrees if the y stepper isn't locked

AFAIK we support all the combinations of cartesian/coreXYZ/markforged for movement axis. you define the relationships between the axis and the drives in M669 as a matrix
https://duet3d.dozuki.com/Wiki/M669#Section_Parameters_for_Cartesian_CoreXY_CoreXZ_CoreXYU_CoreXYUV_MarkForged_kinematics_RRF_2_03_and_later_only

Where this (and other recent threads) differ is adding the extruder into the mix. It gets complicated because of support for pressure advance on the extruder and will get more complex again with input shaping on the movement axis.

JoergS5

@t3p3tony we wrote nearly the same

https://forum.duet3d.com/topic/25470/corexe-will-it-work-any-thoughts-is-it-even-worth-it/10

T3P3Tony

@joergs5 yep!