One fix for the filament sensor stuck "on"  

  RSS
Jeff Lastofka
(@jeff-lastofka)
Active Member

Mine worked right when I first assembled the printer but then stopped working with it being stuck "on" all the time - LCD sensor display shows "1" all the time. A little investigating showed that loosening the screw holding the IR board down and tilting the board so the sensor underneath moved a little away from the filament side made it work. Turns out the mounting surface under the sensor board wasn't flat and I just had to file away a tiny bit of material to make it flat and now it works great. Mine was so close to working that it worked at first before it settled into final position I guess...

It's a thing people can check if they're having sensor trouble like this. The plastic printed part underneath isn't perfect of course and a little build up like I had can make all the difference. I'll attach a photo of where I reached in with a small triangular needle file and just swiped away a little material. Didn't take much at all.

Posted : 30/09/2019 7:12 pm
Michel liked
whiting.peter
(@whiting-peter)
New Member

Same problem. Same observation. My fix (bodge)was to mount the detector with a spacer washer - 0.8mm plasticard.

Posted : 28/10/2019 8:40 pm
rmm200
(@rmm200)
Noble Member

A spacer cut from a bit of a Live Z calibration circle seems to work perfectly.

Posted : 29/10/2019 12:04 am
MGoleta
(@mgoleta)
Active Member

I had similar problem with the MK3S I just got.   Mine turned out to be that the holder for the steel bearing had print residue in it so the ball didn't sit deep enough in the hole.  Removed the ~1mm of filament strand in the hole which corrected the problem.

Posted : 11/11/2019 12:06 am
Jerry liked
Michel
(@michel-2)
Active Member

Same problem here. Adding a spacer did the trick. Filament sensor now works. Thank you!

Repro steps:

  1. Go to Support > Sensors. IR will be stuck on 1 with no filament inserted.
  2. While still looking at the IR value,  remove the sensor. It should change to 0. (if it doesn't you have a different problem)
  3. Screw the sensor back in and note when IR switches to 1.
  4. Add a spacer below the sensor so that it sits high enough to read 0
  5. Insert a small hex key into the filament hole, make sure it switches to 1 and back to 0 when removed.
------------------------------------------
Prusa I3 MK3S...
Posted : 22/11/2019 6:30 pm
Jeff Lastofka
(@jeff-lastofka)
Active Member

It's fun to post something and have someone find it useful and then get some feedback from that:-)

These forums are a great resource.

Posted : 22/11/2019 6:54 pm
Dave
 Dave
(@dave-11)
Eminent Member

Had the same problem after building from a kit. The various workarounds here help, but don't solve the actual problem -- the imprecision of the print of the tiny lever arm that's pushed by the steel ball into the path of the IR sensor.

A solid, permanent fix is simply to break down the extruder to get to the lever, take it out, file the tongue that shifts into the IR path so it's nice and square. Slip it back in and test with the head disassembled, remove and adjust (by filing) as necessary the get 100% reliable operation.

The dimensional accuracy necessary for that tiny part demands printing with something finer than the standard 0.4mm nozzle, 0.3 would probably work, and printing nice and slow. However, that's not how it was printed. I'm thinking of printing a new one as noted to see if I can improve the quality right out of the printer.

This post was modified 2 years ago by Dave
Posted : 04/12/2019 10:44 pm
EricM
(@ericm)
New Member

@chex,  @jeff-lastofka

Hi all, great tips. Combined these to "just loosen the screw holding the IR sensor a bit". Seems I was somewhat over-motivated when fastening it ... Now sensor woks as expected.

Thank you

Posted : 09/12/2019 4:01 pm
sharpie
(@sharpie)
New Member

Hi all,

I had the same issue and did not want to take everything apart again. Here It would be great if the build would include some sort of unit tests before everything is assembled. I ended up shimming the left side (looking at it from the front) of the IR sensor PCB with the thin bit on the end of one of the cable ties.

Unfortunately I did not know back then that you can monitor the sensor signal in the menu, as chex has already written: Support -> Sensor Info -> IR

Posted : 12/12/2019 9:40 am
rmm200
(@rmm200)
Noble Member

Best thing about the sensor - it is easy to get to. Just one screw holds the top cap on.

No need to disassemble the extruder for this problem.

Removing the ball assembly - I am not sure about. That might be more serious.

Posted : 12/12/2019 3:53 pm
mallen
(@mallen)
New Member
Posted by: @david-w107

Had the same problem after building from a kit. The various workarounds here help, but don't solve the actual problem -- the imprecision of the print of the tiny lever arm that's pushed by the steel ball into the path of the IR sensor.

A solid, permanent fix is simply to break down the extruder to get to the lever, take it out, file the tongue that shifts into the IR path so it's nice and square. Slip it back in and test with the head disassembled, remove and adjust (by filing) as necessary the get 100% reliable operation.

The dimensional accuracy necessary for that tiny part demands printing with something finer than the standard 0.4mm nozzle, 0.3 would probably work, and printing nice and slow. However, that's not how it was printed. I'm thinking of printing a new one as noted to see if I can improve the quality right out of the printer.

I agree that the lever/tongue needs filed but I didn't want to take mine apart to get to it. And I was afraid of trying to file it while it was still in the head. So I used a .85mm thick nylon washer under the sensor board and it took care of my problem

Posted : 19/01/2020 7:05 am
Graham A
(@graham-a-2)
Active Member

@david-w107

Just adding this for anyone else searching this topic/thread.

I just re-printed the level using 0.1mm layer height, and it immediately worked perfectly.  Again, nice and slow and the lower layer height seemed to help the tolerances.  

 

Posted : 19/02/2020 4:20 am
Share:

Please Login or Register