Even More Reliable Filament Runout Sensor and Spool Holder by bwaslo

Last crawled date: 3 years ago

This version is operatonally derived from my earlier "Reliable Filament Runout Sensor" (https://www.thingiverse.com/thing:2483919) from last year. Both versions are designed to mount on a 2020 bar which can be part of the printer or part of a separate filament holder. The first design works very reliably when filament runs out and the end of it comes free from the spool. BUT -- there are times when the filament is taped to or bent through an inaccessible hole in the spool so that it can't easily come free. Then the filament doesn't get pulled through the trip sensor. It doesn't happen often, but can be kind of disastrous when it does. Like, when you already have more than a dozen hours into making a big print and then the filament stops flowing as the printer thinks it is still printing and keeps going. One time while I was watching, it pulled the spool right off its rollers (which inspired the tension sensing method for this second version).
So, this version still works (for freely flowing filament) by using a gravity-powered trip arm that is held up by filament (when filament runs out, the arm falls and trips a switch). But this design adds another sensor that activates on high filament tension. Should the filament grab at the end of a spool (or even before), the spool is tugged forward. But the spool holder now can tip on bearings (the 608 skate bearings you still have -- admit it! -- from building in the fidget spinner craze) so that the holder tips forward and taps a mechanical microswitch when filament tension to the spool gets high. Both sensors are mechanical switches now (the standard microswitch end-stop boards, https://folgertech.com/collections/sensors-and-modules/products/end-stop), as they are less expensive and don't have problems with IR light not getting blocked well by some plastic parts, very non-fussy.
The tipping spool holder scheme works quite well as tipping requires less force as the spool empties. The gravity trip arm sensor is balanced (using yet another 608 bearing as a weight) so that the filament feels very little tug from the arm, but once it does run out, the weight goes out of balance and pushes the microswitch with good force. One of the screws mounting the microswitch boards should be longer, to feed through the curved slot in the lever arm. This is to keep the balanced arm from being able to tip the wrong way.
The setup ideally feeds the filament through a "reverse bowden" tube (though it doesn't have to) which completely prevents filament from ever spilling over the edge of the spool --strongly recommended if you have a direct-feed extruder. Either way, use two PC4-M10 Bowden tube bolt-on connectors (Amazon or Aliexpress) to hold and feed the 4mm (2mm internal) PTFE tubing used to guide the filament from the opening of the gravity-trip arm to the printer or other frame. The tube can extend all the way to the extruder for a "reverse bowden".
The spool holder is wide enough to hold any spools up to several kg, without needing any adjustments Spools are simply set on top of the curved rollers-- no rods to feed through, no holder width to adjust, fast and easy to change spools. The spool holder rolllers are tapered (narrower toward center) so that the spool tends to center, helping alignment with the gravity-trip arm. You'll need two 1/4" x 5.5" (or metric equivalent) and nuts to put through the spool holder rollers. Don't be tempted to use lightweight bolts, as (at least at the back) the weight of the bolt is needed to keep the structure from tipping too easily.
The photos show other plastic pieces that were put together for mounting a webcam (on my printer -- an FT-5 -- it turned out to be in the right placement for monitoring printing via Octopi). I didn't include STLs for those parts as they are a different function and would depend on printer and camera types.
[Edit Sept 2018: I managed to break my center guide when a blob on the filament caught in the PTFE tube and I tried to yank it out from the spool end. Since I needed to print a new one, I also strengthened the piece to prevent breaking should I do something dumb like that again. The new piece also has a place for a second PC4-M10, for looping the tube around like I have in the pictures -- I'd recommend using the on named "Center_Guide_redone" rather than "Center_Guide_old...."]
Attached also is a drawing for how to wire the end stop boards, along with code changes for Marlin so that it can use the sensors and do it's cool alarm, pause, and wait for filament change trick. You'll need 8 608 skate bearings, use the smoothest rolling ones for the spool rollers, and the least rolly one of course for the balance weight.