Apologies for delay on IR sensor support. I finally received the backorder kit from Prusa. I will get it out asap once I can test. In the meantime, try this.
First official release is in progress! Finally getting out of beta. Less than one year. ;-) I am still making updates/improvements so you may want to add "watch" to this project. Example is new hybrid sensor mod I released.
New instructions: I moved these to their (hopefully) permanent home. The transfer was manual so let me know if you see any issues.
I wanted to see if I could make a lighter weight extruder to unlock more speed on the MK3. I also wanted to try some new innovations, fix some nagging issues, easier adjustments, etc. while keeping the good bits from PR. I started with a blank piece of paper, solving the problems I saw and ended up with something related to e3d Titan and Bondtech BMG. But it’s tailored for Prusa MK3, includes more features and you can make it yourself for much less.
Remove: 200g of moving mass and center motor on belt-line to increase acceleration potential
Keep: build volume, e3d v6, Bondtech, laser sensor, other goodness from PR
Improve cooling and nozzle visibility
Increase ease of use -> see Usage
Optimize parts for FDM printing -> RepRap spirit
Purchase as little new hw as possible
- lightweight motor: $8-15 (or use stock for free)
- 5mm shaft: $2-8
- belt loop: $1
- 3 bearings: $2
Main highlights (so far):
- Almost 200g less mass, centered on belt line
- Increased print volume for V6; Or use Volcano with no loss in print volume
- Filament sensor works with any color filament including clear (solved here first)
- Shorter stack and free path to better handle flexible filaments
- Tool-free rack-and-pinion PINDA mount adjustment
- Tool-free quick-lock idler with tension adjust and manual load/unload
- Improved part cooling from both sides
- Full visibility of nozzle printing
- Garter screw to keep V6 sock in place
Supports (now or soon):
E3d V6 or E3d volcano, also Mosquito, Beta
MMU 2.0 (also see Hairlock ) or other multi-filament
Direct filament or reverse Bowden
NEMA17 or NEMA14 (deprecated but ask)
Direct or indirect laser filament sensor
Std inlet or full PTFE
Other things that are better:
- Bondtech gear meshing better aligned with taller more central pivot
- Improved torque for pushing filament
- decoupled motor so no heat transfer to filament
- gearing + belt reduces stepper noise into print surfaces
- easier X belt access for maintenance
- larger, less restrictive path for Noctua cooling
Multiple gear ratios supported. More on the way. Note 56/16 is most common.
Short path anim shows how the filament has a short path from the cartridge to the inlet while traveling between the Bondtech gears. This better handles flexible filaments as there is no room to deviate. Also the motor is not attached to either gear so there is no heat transfer. Witness Note video shows older design, but performance is the same
No Firmware change required! That said, there are things you may want to change for convenience. See assembly instructions. Indirect filament sensor does require a small change which is very simple.
Update 6/1/19: I added step files for version r1b. I may make some notable changes to accommodate some things I've been wanting to do, like mosquito support. That may involve removing the key for example. So folks can hack up the step to continue support for the current version and add mods for other platforms, etc. Please link a remix back to this project so people can find. Enjoy!
Update 3/24/19: finally updated the omega coolers to R1 as they had some fitment issues. Made a hair thicker for folks using exotic filaments or 0.6 nozzles.
Update 2/2/19: added new hybrid filament sensor (beta). This requires no firmware changes!. The sensor simultaneously reads the filament and bearing. Initial testing looks promising. It is definitely better than stock direct sensor. I hope it achieves 100% reliability like my indirect solution. This is nicer too because it sits recessed and can use the standard covers. To build, print Bcart(ridge) and Bplate in PLA 0.15 layer. Note that you will see screw hole punch through which is fine. Screw sensor into cartridge using a button head MK3x8 or 10mm. I found that MK2.5 button head also works. I only mention because I found one of those oddly in my spare bag from Prusa. Place 2 673 bearings on plate spindles. Now snap the plate into the side of the cartridge and install cartridge in normal position. Note: if you flashed firmware for indirect sensor (Y flip), you need to undo that because this does not need that modification. Please let me know how if you find a filament not reliable.
Update 1/31/19: increased cavity on inlet to fit collet clip if you use one. small tweak to cartridge to improve printing.
Update 1/13/19: first official release. Also multiple gear ratios supported. Note, if you are upgrading from B4 or newer, the Ecages and Prack have changed. Other parts, including carriage, spider, inlet, cartridge, etc should still fit.
update 12/12: fixed P rack to print cleaner in PETG and better fit with Ecage B (why nobody complained?)
Update 11/25: added cage option for e3d Beta heatsink with thread mount instead of groove mount.
Also updated carriage and X clamp to offer multiple belt paths, per a million requests. Now supports Bear, or stock with 16T (flat), or stock with larger idler (offset). Recommend cutting 2 10mm piece of belt and insert in paths not used to add bite. For stock idler (20/22T) screw on jammer using a M3x8mm to align belt if desired.
Update 11/13: updated cage to b5 with some more tweaks after testing. I included a second screw hole for holding down the I_cartridge. This is needed for bowden setups like MMU2. I also added a jam block you can screw onto carriage if you want the belt level lower if using smaller X idler pulley. Beefed up some areas including idler after long term testing. I recommend PETG for most parts now, though PLA is also viable.
update 11/3: added a modified frame bracket for MMU2 that gives a bit more clearance for wiring on the tail/fin. I also rotated it forward 15 deg to assist clearance but also help the FINDA ball settle with a bit more force. Probably negligible.
update 11/1: I uploaded the new b5 cage for testing. The idea is that cage S can be printed in a high temp friendly filament, like PC, ABS, etc. Then the rest of the cage can print in more print friendly material like PLA or PETG since it does not see the high heat. The assembly is a bit trickier as you have to slip the new cage S over the heatsink first and then assembly the rest of the cage. Leave it loose on the heat block while you assemble cage and then snap it in place in front right corner. Use 2 M3x18mm screws for back Noctua fan and 2 M2x12mm for front holes. These help secure the cage S too.
update 10/19: updated Omega to improve blower fit. The CAD model I have is the wrong fillet radius. Also, if you have an older wheel, replace with b5. My 4 month old wheel cracked so I reinforced that area.
In other news, I have MMU2 mostly working now and it looks like I may need no changes over what is already here. Stay tuned.
Update 10/16: added OmegaV for Volcano. see below for tricky print. I am working on other b5 changes to Ecage.
Update 10/15: I received my MMU2, so I am testing to get everything working with skelestruder. Also added tail step file to people could make cable chain adapters, etc.
Meanwhile, I designed a new omega cooler. This is experimental and needs new spider and shroud. I am still testing so be warned. For V6 only, working on Volcano version. Use M3x12 to screw spider to cage without bracket (not used now). Instead used M3x10 to screw omega to spider while also fitting hook in shroud. Then screw blower onto spider with M3x18. You can use hex or Nyloc nuts. Printing: omega is challenging, needs 4mm brim + support base only with 40 degree threshold using 0.15 layer height. Let me know how it goes.
Update 9/15: fixed rotation on Ecage F17. Removed nema14 support temporarily. PM if you need it.
Update 9/8: add direct cover that will connect to MMU2. need brass nut, see instructions
Update 7/12: finally got the indirect filament sensor added. Assembly instructions updated. Makes filament sensor work with any color filament, even clear. Fixed small issue on I-lid b2a.
Note 6/30: if you build b2 or newer with a v6 as opposed to volcano, you may fail XYZ cal because you have more Z range than the firmware is expecting. You can change #define Z_MAX_POS 220 in Configuration_prusa.h, OR temporarily position the PINDA below the nozzle, run the XYZ cal, then return the PINDA to normal position, run Z only cal, then run First Layer Cal.
Update 6/2: finally added fin part as alternative to tail, for testing. Do this to your stock machine: unhook X belt so carriage can slide freely. Slide it from one side to the other and notice the high tension from the umbilical connection. I am trying another route following the ancient philosophy of "bend don't break" and inspired by cartilage in shark fins. I built in the strain relief into the part so it stays light and keeps wires stable. Seems to work so far.
Updated assembly instructions.