Loading
Ludwig3D

Filament Runout or Jam Sensor

by Ludwig3D Aug 28, 2018
Download All Files

Thing Apps Enabled

Please Login to Comment

There are times when the filament is stuck and stepper starts skipping steps and it kind of jumps back a little. I assume that in such case the slotted wheel will still provide back feed?

Hi Glassy
Good point. I haven't encountered this problem though.
My Tarantula has a Bowden setup and there is hopefully enough drag in the tubing and the channel of the device to stop this happening. But if it had stopped just on the cusp of the wheel moving from light to dark or vice versa it could register movement still maybe. It would also depend on whether the jumping of the filament occurred frequently enough for a jam not to register.

On direct extruder i.e. on Raise3D printer when filament gets stuck it kinda "clicks" and there is a small constant forward/backward movement.
Otherwise my sincere congratulations! Very nice solution.

Thanks Glassy!
The sensor will detect around 0.3mm of filament movement so if it is jiggling continuously more than this without any pauses in the movement it may not be detected by the sensor. But it only has to be stopped or slowed down briefly, once, for the sensor to detect an error so if the jiggling doesn't start straight away when there is an obstruction or if it not continuous then I think the sensor would work. Once an error signal has been sent to the mainboard my Tevo Tarantula with Marlin pauses, even if the error signal is only momentary and clears again. I presume marlin would work the same on other printers it is installed on but I cannot verify this. I suppose if this was a problem one solution would be to make the slots in the disk wider so the filament had to be moved more for movement to be detected. One or two resistor values in the circuit could be changed to accommodate the longer time between movement detections but the downside would be that the sensor would respond more slowly to error events.
When I get the time I might see if I can reproduce this behaviour on my Tarantula.....

So, I have tried to reproduce the backward/forward movement when there is a filament jam on my Tevo Tarantula with a Bowden extruder with some success.
If I emulate a blocked nozzle by turning the extruder temp down to just above the min extruder temp on my machine (170C) and using filament that melts at a higher temp (eg ABS) the jam detection works fine and an error is picked up as the filament movement initially slows as it backs up in the tubing then after about 10 seconds it starts to flick back and forth every second or two as the extruder gear eats into the filament. The error light still comes on during this time so this does not stop the jam detection working in my set up - when printing the print will pause after any single error detection, and it still detects errors intermittently when the filament is flicking. I also tried a flexible filament (Hobby King) and this worked too although for flexibles I have a short length of PTFE tubing with a V shaped end added to my extruder above the gear to stop the filament kinking out of the extruder and with this modification the filament detector behaved similarly to when using rigid filament.
Unfortunately I don't have a printer with a direct drive extruder to test this on.

I came up with a way to imitate a blocked direct drive extruder - I replaced the PTFE tube that connects my Bowden extruder to my hot end with a shorter kinked off (to simulate a blockage) length of PTFE tube inserted into the outlet of my extruder. I started with a length of 15cm and tested at 1cm shorter lengths all the way down to zero. The jam detector worked ok at each length and detected the jam.
There was always a brief time of a second or two between when the filament came to a halt and any jittering movement started and an error was always detected at this point. Errors were also detected at least every few seconds as the filament jittered, usually more frequently.
Interestingly the jittering became less as the length of tube shortened and the blockage occurred closer to the extruder gear. I couldn't get the filament to jitter once the distance from the gear to the blockage was less than about 5cm.
So based on this testing I think the detector should work ok with a direct drive extruder, but I have not actually tested it on one. What if the extruder gear had particularly coarse teeth for instance? I don't think they vary that much....
If anyone has any problems with the detector on their set up could they please post a comment here.