Toolchanger for CoreXY printers

One of the issues of the usual dual extrusions implementations, IMHO, is that that the unused nozzle(s) are a steady source print hazels and failures, either by damaging the printed object (unequal Z height) or nozzle clogging if materials with different melting temperatures are used (ABS and PLA) or simply oozing. Recently some workarounds appeared. 2-in 1-out hotends (eg. E3D Cyclops) solve most of the dual extrusion issues (nozzle alignment, oozing), but are limited to materials with similar melting point. Printers that lift the unused nozzle (eg. Ultimaker 3) are another approach, but I suspect that mechanical constraints (size, weight) are limiting this solution to 2 or 4 nozzles.

Tool changing is not a new idea, but in my opinion, the way forward with regards to multi extrusion. It allows different extrusion temperatures (multi material), and scales beyond 2-4 extrudes. Having multiple extruders available is already and advantage by itself, without the need for multi extrusion prints, as it allows you to keep your eg. filaments loaded, use hotends with only one type of material (reduces the chances of nozzle clogging) or quickly use a different nozzle diameter without screwing around (sorry for the bad pun).

The CoreXY gantry is well suited for this purpose as the space between the two motors can be utilized to store several extrudes that are changeable without further motors/servos etc. I've upload a video from my prototype to illustrate it (with only one instead of three hotends, but you get the idea).

[www.youtube.com]
[www.youtube.com]

I used a microscope camera and 1mm grid paper to check for repeatability over several insertions/releases of this simple mechanism and came to a value better than 50µm, far below the typical nozzle diameter and therefor good enough for the purpose. The coupling is made of 4mm pins and sintered bronze bushing.



I still need to finish the Z Axis of my prototype printer, so it will take a couple of weeks before I'm able to test this idea in a complete environment, but I thought it would be worthwhile to write about.

Very interesting design, it does seem to be an excellent solution to dual extrusion. I would love to see more pictures of the designs and more detail of the mechanics of the system. I am curious, though how do you hope to address possible tangling of the wires and do you have any ideas for further improving the accuracy? Best of luck and I can't wait to see it work.

Sorry that it took me that long to come back to this.

Quote
Trexation
I would love to see more pictures of the designs and more detail of the mechanics of the system.

I just uploaded another video, from a birds view perspective. [www.youtube.com]
The principle is very simple, the hotend mount has two prisms at the bottom and top, that used to lock it in the parking position by moving the hotend back -> left -> front. Reversing this sequence, releases the hotend. A spring like element pushes the hotend to the front to secure it while parked (new since last post). So far I did about 1500 park/unpark cycles, without a single failure.

The three parts that make the parking holder:

Quote
Trexation
I am curious, though how do you hope to address possible tangling of the wires and do you have any ideas for further improving the accuracy?

Actually, I don't think that tangling of the wires will be a big issue. Hotends are always parked/unparked at the same positions, so it is impossible to tie a knot.

With regards to accuracy; I used a USB microscope, OpenCV and some Python script to get measurements for the repeatability of the nozzle position after parking and unparking. It seems to be at most 10µm (2 pixels) in any direction.
This is an image of the nozzle taken from the microscope . The nozzle size is 0.4mm and the diameter of the green circle is 92 pixels, so the resolution is about 5µm/pixel.

I case someone wants to try it, I attached STL's for the hotend mount and the parking holder. I used PET-G for my prototypes. I think that PLA is to stiff and brittle, ABS might be an alternative. The spacing of the pins center-to-center is 29mm.

--
Roland

Edited 1 time(s). Last edit at 03/24/2017 01:18PM by rmie.

How is the project coming along? repeatability looks really good.

Quote
Evan
How is the project coming along? repeatability looks really good.

Much slower than I had hoped for. The first major milestone was reached last week, dual extrusion prints: [youtu.be]

I started a project on hackaday.io yesterday, and have plans to publish a lot of the details over the coming days.

[hackaday.io]

That's freakin' cool bro!

Amazing videos! The toolchange looks very fluent.
I wonder how you make sure the z-height is the same on all hotends? Do you set each z-height in the tool change macro?

Quote
o_lampe
I wonder how you make sure the z-height is the same on all hotends? Do you set each z-height in the tool change macro?

I correct for the offset in software. One tool is set to zero for bed leveling. Find details about offset calibration here: [hackaday.io]

A discussion about the options for offset correction will be available soon.

A lot changed since the last update. Oozing was a major issue, but this is now under control by covering the nozzles of unused/parked hotends with silicon wipers (molded from 350°C silicone). The print head is now supported on four steel pins and printed (instead of brass) bushings. Everything except for steel pins and some aluminium T extrusion is printable.

[hackaday.io]

[www.youtube.com]

The design files, a plugin for Cura to get the tool change into the G-Code and (not yet all) instructions how to build the parts are all published here:

Files
Instructions

The design is highly modular. Integration into other machines boils down to
- redesign the carriage with pins in correct position
- find a way to mount the tool rest

I'm happy to help everyone who wants to try as much as I can.

This is really nice and perfect use case for RAMPS expansion board http://forums.reprap.org/read.php?2,784601

I just visited your hackaday page and read about the USB-camera problem. I have a USB "Endoskop" camera which is much smaller and comes with integrated LEDs too. The focal length is pretty short, but not as short as a microscope. Just thought, I mention it here...

Quote
o_lampe
I just visited your hackaday page and read about the USB-camera problem. I have a USB "Endoskop" camera which is much smaller and comes with integrated LEDs too. The focal length is pretty short, but not as short as a microscope. Just thought, I mention it here...

thanks. I, got a budget ($10) USB endoscope a weeks ago, because of the form factor and the integrated LEDs, but abandoned it for two reasons:
- fix focus 4cm - infinity, the resolution is too low.
- a makro lens in front gives better resolution but a lot of stray light too (light from the LED reflected on lens surfaces), to much blur

Edited 1 time(s). Last edit at 09/16/2017 04:54AM by rmie.