When Elegoo first jebaited us by retconning the promised multi-material system for the Centauri Carbon, I started thinking about what could be done to bring multi-material to the CC1. Unfortunately at the time the OpenCentauri project had not yet decoded the toolhead or heatbed protocols, so it wasn't a practical time to be making any plans. As of recently though, the OpenCentauri project finally finished figuring out all the stock part protocols and Klipper converted the CC1 which granted it compatibility with the Anycubic ACE multi-material system. Elegoo also followed through with their initial promise by releasing the CANVAS kit for the CC1 at a significant discount, but I still think there is a better option. Specifically, a toolchanger.
Last year that pesky patent that was blocking fixed gantry 3D printer toolchangers from being developed without expensive ass licencing finally expired, so now everyone and their dog is rushing to make one. Unfortunately, all the projects that I can find are either for DIY printers like Vorons and DuEnders, unavailable for sale due to licencing exclusivity periods (INDX), or are seeking a million bucks on Kickscammer to maybe deliver some units to backers before going bankrupt (looking at you Wondermaker you fucking bums
). Naturally I shall be channelling my inner pirate to steal everything from all of them to make my own.
Enter stage: Centauri Changer (name subject to change)
A hybrid system (read: mutant freak) of cherry picked reverse engineered aspects from 7 different toolchangers, Centauri Changer will bring true poop-less multi-material printing to the CC1 for a hopefully worthwhile price.
My main goal is to blow up:
The Centauri Carbon is a very affordable machine, so I want to make this modification affordable too. Unfortunately, the stock mainboard does not have enough overhead to run the needed Klipper plugins or enough USB inputs for needed modules, so a new 32bit mainboard + TMC drivers + a Klipper compatible SBC will be required from the get go. An extra ~200 watts of 24v power will also be required to support pre-heating multiple hotends, and a new touchscreen too since the stock one is directly controlled by the stock MCU and I don't think I have the skill to learn SPI and reverse engineer the display protocol from nothing in 3 months. This much starting cost is considered unavoidable. Fortunately, I think the stock toolhead board will be usable. The COSMOS (Centauri Carbon Klipper fork) page says the part cooling fan's tach pin can be repurposed for things like a filament detector, so I can probably use that as my toolhead side tool sensor. It was also originally intended to have under-tool addressable LED lighting, and the traces for that still exist on the board and can be easily enabled and used with a few wires.
Currently the plan for the mechanism itself is a simplified derivative of Srin's unnamed and unreleased toolchanger which is itself a simplified derivative of Irbis3D's Medusa-HC project. The tool will be be held in the toolhead by 4mm pin & bushings for X and Z securement, and magnets + filament gear tension for Y. Technically the Medusa-HC design could just be put straight into a CC, but the tool and dock width would hard cap the maximum tools at 4 and sacrifice up to a third of the Y axis build dimension. The Medusa-HC also requires a 30mm 24v heatbreak fan per tool, which I hope to eliminate the need for as the good ones cost a fair bit per unit.
As it currently looks in my head, each tool will consist of only a hotend on an aluminium mounting plate with a printed filament path and essential 4 wire umbilical going to the 32bit mainboard. The toolhead electronics, extruder motor & gears, part cooling fans, and heatbreak fan will remain fixed on the toolhead. The only functional part of the stock toolhead that will not be re-used will be the hotend. The CC's stock hotend has only the thin heatbreak tube to connect the block to the heatsink, which is far too weak to be used with a physical nozzle offset probe. Over time it would wiggle back and forth and weaken the press fit enough to leak molten filament up into the toolhead. I am also likely to have some nozzle crashes while tuning this thing, so I will be opting for a stronger 3rd party hotend that facilitates easy attachment to the mounting plate.
I will provide more detailed explanation on how the system is to physically function when I have some non-napkin diagrams and/or some CAD progress to share.
* Parts will be designed such that they can be cut, finished, and assembled with non-specialty power tools wherever that is reasonably achievable. An electric drill will be needed, but other power tools like dremels, grinders, drill presses, or belt sanders will merely make assembly easier.
** Using only metal is impossible without CNC machined parts. Some bits will have to be printed by nature of the no-CNC rule, so I am classifying "full metal" as "metal will be used everywhere possible".
Last year that pesky patent that was blocking fixed gantry 3D printer toolchangers from being developed without expensive ass licencing finally expired, so now everyone and their dog is rushing to make one. Unfortunately, all the projects that I can find are either for DIY printers like Vorons and DuEnders, unavailable for sale due to licencing exclusivity periods (INDX), or are seeking a million bucks on Kickscammer to maybe deliver some units to backers before going bankrupt (looking at you Wondermaker you fucking bums
). Naturally I shall be channelling my inner pirate to steal everything from all of them to make my own.Enter stage: Centauri Changer (name subject to change)
A hybrid system (read: mutant freak) of cherry picked reverse engineered aspects from 7 different toolchangers, Centauri Changer will bring true poop-less multi-material printing to the CC1 for a hopefully worthwhile price.
My main goal is to blow up:
- Fit 4 tools inside
- Individual temperature control for each tool including pre-heating for swift changes
- Modular dock system with active heatbreak cooling and docking sensors for failure detection
- Maxwell Coupling for 0 drift tool mating
- Automatic nozzle cleaning
- Automatic one button XYZ offset calibration
- Revise the motion system or bed float to retain maximal print volume
- Recirculating AUX part cooling fan
- Use globally sourceable off-the-shelf parts where possible
- Design parts to reduce need for power tools*
- Re-use stock components wherever possible
- Robust pre-flight sanity checks to prevent situations like running hot chamber jobs when PLA is loaded
- Minimal structural modifications to the Centauri Carbon chassis so it resembles a standard unit from the outside as much as possible
- Make the tools small enough to fit more than 4 in one machine
- Full metal** toolhead construction without use of CNC
- Active chamber heating with thermal runaway detection and fire suppression
- Tri-motor bed tilt levelling
- Assisted filament loading/unloading
The Centauri Carbon is a very affordable machine, so I want to make this modification affordable too. Unfortunately, the stock mainboard does not have enough overhead to run the needed Klipper plugins or enough USB inputs for needed modules, so a new 32bit mainboard + TMC drivers + a Klipper compatible SBC will be required from the get go. An extra ~200 watts of 24v power will also be required to support pre-heating multiple hotends, and a new touchscreen too since the stock one is directly controlled by the stock MCU and I don't think I have the skill to learn SPI and reverse engineer the display protocol from nothing in 3 months. This much starting cost is considered unavoidable. Fortunately, I think the stock toolhead board will be usable. The COSMOS (Centauri Carbon Klipper fork) page says the part cooling fan's tach pin can be repurposed for things like a filament detector, so I can probably use that as my toolhead side tool sensor. It was also originally intended to have under-tool addressable LED lighting, and the traces for that still exist on the board and can be easily enabled and used with a few wires.
Currently the plan for the mechanism itself is a simplified derivative of Srin's unnamed and unreleased toolchanger which is itself a simplified derivative of Irbis3D's Medusa-HC project. The tool will be be held in the toolhead by 4mm pin & bushings for X and Z securement, and magnets + filament gear tension for Y. Technically the Medusa-HC design could just be put straight into a CC, but the tool and dock width would hard cap the maximum tools at 4 and sacrifice up to a third of the Y axis build dimension. The Medusa-HC also requires a 30mm 24v heatbreak fan per tool, which I hope to eliminate the need for as the good ones cost a fair bit per unit.
As it currently looks in my head, each tool will consist of only a hotend on an aluminium mounting plate with a printed filament path and essential 4 wire umbilical going to the 32bit mainboard. The toolhead electronics, extruder motor & gears, part cooling fans, and heatbreak fan will remain fixed on the toolhead. The only functional part of the stock toolhead that will not be re-used will be the hotend. The CC's stock hotend has only the thin heatbreak tube to connect the block to the heatsink, which is far too weak to be used with a physical nozzle offset probe. Over time it would wiggle back and forth and weaken the press fit enough to leak molten filament up into the toolhead. I am also likely to have some nozzle crashes while tuning this thing, so I will be opting for a stronger 3rd party hotend that facilitates easy attachment to the mounting plate.
I will provide more detailed explanation on how the system is to physically function when I have some non-napkin diagrams and/or some CAD progress to share.
* Parts will be designed such that they can be cut, finished, and assembled with non-specialty power tools wherever that is reasonably achievable. An electric drill will be needed, but other power tools like dremels, grinders, drill presses, or belt sanders will merely make assembly easier.
** Using only metal is impossible without CNC machined parts. Some bits will have to be printed by nature of the no-CNC rule, so I am classifying "full metal" as "metal will be used everywhere possible".
