Prusa Mk3 indirect filament sensor for all filaments (transparent, black, shiny, whatever).

thingiverse

I think the Prusa Mk3 filament sensor is a fantastic innovation. However, some filament types can be tricky for the sensor to detect. Various attempts have been made to develop indirect sensors that rely on sensing a proxy (usually a bearing) moving with the filament. The filament sensor is optimized for reading stainless steel surfaces, making a ball bearing an ideal choice. Nevertheless, many of these require firmware modifications (https://www.thingiverse.com/thing:3091625, https://www.thingiverse.com/thing:3178706), or necessitate a separate housing, leaving a void where the old sensor was (https://www.thingiverse.com/thing:3129921). For my design, I aimed to achieve several key objectives: • To significantly improve the reliability of the filament sensor. As the MMU will undoubtedly be linked to the filament sensor in the future, I wanted my filament sensor to be extremely reliable regardless of the filament. • To enhance the reliability of the feed to and from the extruder. This means a clean, tightly constrained pathway, with no voids or sharp edges (such as those found in the existing sensor chamber). • To avoid firmware modifications altogether. • To steer clear of any other negative compromises in the extruder. For instance, access to the bondtech door had to be preserved. • To work seamlessly with existing Prusa parts, such as the top cover (as there are numerous useful modifications of that). The X-carriage plate had to remain unchanged (who would want to disassemble their x-axis if they can avoid it). This design meets all those goals and requires only two printed parts: a single MF63zz bearing and a 10mm countersunk M3 bolt. The parts can be successfully printed using standard Prusa profiles. I employed the '0.15 - Optimal Mk3' profile, with the shells bumped up to three, and gyroid infill (it's truly awesome). To achieve this, I rotated the bearing so that the filament sensor read the side of the flange. Using a printed wedge with a window cut in it ensured that the sensor couldn't see anything else, such as other non-rotating parts of the bearing. This design also means that the whole system is well secured and robust. As a result, I didn't have to move the filament sensor much at all, so the sensor wire will still pass through the existing x-carriage, and the original top plate screw holes are in their correct positions. To install: • Print the parts using the '0.15 Optimal Mk3' profile with 3 shells and gyroid infill. Use the standard Prusa orientation for the extruder body. • Reinstall the extruder body. Check that autoload and runout detection work, and check the extruder menu under support. If there's no print running it should report intensity. Cool benefits: • You can probably print the extruder body in any color you like, so long as the wedge is black. The bearing and the wedge prevent the sensor from seeing much of anything else. • The filament path is now extremely constrained. There's no chance of a bit of rogue filament spearing off into the sensor chamber and killing your filament sensor. Another useful mod: For even more reliability, try this mod so you don't need to little bit of PTFE tube. Instead, you can run the main PTFE tube all the way to the filament sensor chamber. https://www.thingiverse.com/thing:3129388 Installation video: https://www.youtube.com/watch?v=Tj9WczUMouc Update 20/11/18: Added 3 .STEP files for people to play with. The extruder body and wedge, and the boolean shape used to cut the bearing chamber.