Improved Inductive Filament Runout Sensor/Filament Guide

In the never ending quest for fuller automation and smoother filament feed, I designed this to solve some issues.  I recently revisited automated filament runout handling (Marlin 1.1.9+), as well as smoother filament feeding to avoid the “starved extruder” condition.  I have an original i3 Rework with a black acrylic frame.  Today I know this as an MK2, though mine is modified and refactored to include an aluminum build plate for better heat distribution and PEI sheet for improved adhesion without glue/tape, etc.  I can also fine-adjust the BL Touch Probe through the use of spring loading and thumbscrews, but I digress…

I feed my filament directly out of an EIBOS Easdry single roll filament dryer.  It sits NEXT to my printer for my setup not above it or below it and I run a Titan Aero extruder, and interested in smooth feeding as well as being able to detect and handle the filament runout condition. After having switched to Prusa Slicer and learning I could insert G-codes to park the nozzle and switch colors I was one step closer to handling the filament runout condition.

I had tried this before with a mechanical “roller microswitch” but Marlin 1.1.9 introduced new handling of end stops which made them more susceptible to noise.  At that time (2016), the 3D printing world was quite a different ecosystem and certainly not as diverse with available information as today (2023).  I backed out the filament runout code because I couldn't even home the machine without spurious resets.  I later learned that adding a 0.1 uf cap across my end stops addressed the noise susceptibility but other priorities had me set aside 3D printing until recently.

This design utilizes an old PINDA v2 I had (I don't use the thermistor) and it has an active low output with a passive pull-up on the RAMPS 1.4 board in my system.  I wanted some adjustability in the sensitivity that could be done without disconnecting the wire (to avoid twisting), as well as provide a nice arc filament path between my dryer and the extruder that fed with as little friction as possible.

This design was made as a single piece in openSCAD, and then cut in half using Prusa Slicer's cut operator.  This exposed the internal cavities face down on the print bed for a nice clean surface.  Since this is not subject to mechanical stress and strain other than filament passing through, the two halves after printing are reassembled into one with small zip ties which are sufficient and non-ferrous so as not to disturb the flux sensing field of the sensor.  A 6mm steel bearing sits in the cavity and is held up by the filament, that is until it runs out, tripping the sensor to its opposite state and invoking the filament runout handling code.

If you notice from the photos, small lengths of PTFE are used as smooth feeding ‘bearing’ material where the filament leaves my dryer, enters and departs the sensor, and into the entry of my extruder.  The entry and exit are cone shaped to allow some flexing of the PTFE in the sensor, and there is a PTFE channel upon entry and exit with proper trench alignment for the filament path inside.  The PTFE channels slightly intersect the lower zip tie holes and when tightened serve to capture the PTFE so it stays put.

One of the nuts (in the window), is used to adjust the height of the sensor, and the one on top locks the adjustment into place.  A mount is provided that allows the sensor to be attached to the top of the Z-frame just below the top mounts for the rod and lead screws.  Bracing on the mount increases its stiffness, and the arc and slight flexibility of the PTFE tube in various parts of the filament path (see photos) actually flexes slightly buffering the demand/supply surges between the spool and the extruder that occur during extreme moves of the X-carriage.

Printed in Hatchbox PLA Pro+ at 0.2mm resolution.  Note: the modification to the EIBOS dryer is not my own design but one of many for PTFE departure from a dry box.  It is only included in the photos to go with the above explanation and help complete the illustration

UPDATE: This design has been tested with both the PINDA V2 and SUPER PINDA sensors and works satisfactorily and without any false alarms.  My first attempt used a steel BB, but they are too small and don't reliably stay over the filament trench.  The 6mm steel bearing does the job satisfactorily.