Constant failure to pass filament through the toolhead sensor

Bambu Lab X1 Series Bambu Lab X1C
1

I have been fighting for over a year with this. I’m starting to develop anxiety when i hear the ams motors.

This printer is constantly, eternally, fucking around with the filament and trying to feed it, unfeed it, feed it, unfeed it. cant get it through the filament sensor, cant unfeed it because it hasnt even attempted to cut the filament so of course it cant unfeed it. its like a child crying because it wont let go of an animal stinging it. so it ends up grinding away at the filament and i have to cut off a meter of the stuff to get rid of the divot its put in there and cant feed past anymore.

What happened to my lovely reliable printer. I’m about to throw this thing through a window.

How do you get this fucking thing to have a reliable filament path? i’ve done everything. i’ve tried any number of addons to control the bowden tube paths, got the support sanctioned mods inside the ams, tried a bunch of different mods to the bowden where it feeds into the toolhead. I’ve replaced every tube everywhere repeatedly. I’ve replaced the feeders in my ams. i’ve replaced the filament sensor. I’m losing my mind here. I just want the thing to be able to load filament without conducting a seance.

2

Key for me was to:

  • Have a sharp and precise 90° clean cut at the PTFE ends (both in and outsides the AMS)
  • Ensure the PTFE’s are pushed all the way in (both in and outsides the AMS)
  • Not to install too many mods (tried early on to extend the PTFE path: not a good idea at all)
  • Key: Make the shortest possible PTFE path (my AMS sit to the left of the printer with the Hub hanging in the air)
  • Key: Install AMS savers in the AMS (the ones I use were probably ripped as they were taken down from MakerWorld but these look like they could be the original: Filament guide for AMS saver V2 by madizmo - MakerWorld). This type of AMS saver does not fully eliminate issues but I have found it to help a lot for both winding and unwinding.
1 Like
3
  • I have sharp clean ptfe cuts. The filament sensor design is fundamentally flawed. Literally have little steps cut in it for filament to catch on.
  • PTFE is pushed all the way in
  • The printer is basically stock. I’ve tried the mods, they havent worked, they get removed.
  • I have savers over the top of the inlets yes.

None of this solves the problems:

  1. the filament sensor design is wrong and catches on the filament.
  2. the machine tries to retract filament without cutting the filament or reversing the extruder.
  3. while doing both these things it drives the ams motor grinding a divot in the filament, causing guaranteed future failure.
4

Slight misalignment of the PTFE tube that feeds in to the extruder can prevent filament from actually reaching the sensor. The AMS loads/unloads/loads/unloads until it quits. Try removing and re-seating the PTFE tube.

1 Like
5

dude. constant problem for more than a year. i have tried turning it off and on again, as you say.

6

I have not had a problem with the filament sensor catching and have not seen such issues in the forum. Did you try replacing the filament sensor?
I could well imagine that the coupling at the sensor or the sensor itself is a lemon. I had that kind of bad luck with the LIDAR. Had to replace the printer in the end (but I that was a particular (un-)fortunate case).

Does it try to cut and fails or does it not even try? If it does not even try, a factory reset and update cycle as well as re-installation of your slicer may be needed.
If it does try, I assume you have replaced the cutter blade. Did you check and clean the nozzle and extruder at that time? During a particularly bad extruder-nozzle travelling clog I had my cutter blade being stuck. That was a symptom though rather than a cause.

Yeah, I have had that happen during aforementioned multi-clog. Needed to clean the Extruder and the Nozzle before trying again. In particularly bad cases, for example using carbon filled filament, I can imagine extruder damage as well. In particular if it happens again and again.

7

Heres a video of talking about the design flaws of the filament sensor:

and heres a recording of the printer gcode totally failing to do the correct procedure to load/unload filament. which, notably, it had only just loaded the filament to perform a print, then immediately went to unload the filament again, and failed. why for fucks sake.

It moves about but does not perform the actual cutting action. See the video.

I have checked the blade and disassembled the extruder and everything is totally fine.

It does this at the ams, not the extruder, because its not even getting to the extruder most of the time.

8

according to: AMS keeps loading and unloading | Bambu Lab Wiki

all cables have continuity. the spring isnt caught. the buffer pins read:

gnd > buff_a = 1.0v
gnd > buff_b = 2.2v

buff_b slightly lower than the specced 2.3v. could be my multimeter.

I pulled the blade from the cutter and replaced it with a brand new one, theres no discernable difference in sharpness.

10

Thanks for that video. You have gone into a nice bit of detail on the sensor.

Having checked my spare (my active one does not have any issues), the PTFE connector does have some play but nowhere near as much as shown in your video. The installed one does not have any play in my X1, so together with the z-sliding tube guide bracket above it (did you install a mod there?), I do not get any out-of-plane connector movement.

So it looks like it is triggering the broken/run-out filament logic. Otherwise, it would try to cut.
Interestingly, the filament loading path indication shows a completely empty path. Not even the AMS Hub/Buffer is shown as being loaded. Quite puzzling.

Out of curiosity, does it print OK from the rear spool once past the entry of the filament sensor?

So initially I followed your reasoning that the Filament sensor had a unique issue. Not just on the entry side, but also in its actual filament detection.
With the Video showing an empty filament guide path, I am not so sure anymore though.

I am afraid that at this stage I can only recommend opening a ticket with Bambu. That can take a while, but at least they can read the logs. It can be a frustrating process though.

In the meantime, a factory resets and a SD card replacement may or may not help (I am always flabberghasted when a logic problem turns out to actually be solvable by a SD card replacement :flushed: Not many of those, but they do exist).

11

I recently bought another ams.

I’m still having these problems. its clearly two separate issues, but i suspect the same root cause.

  1. The filament being fed into the tool head is not detected.
  2. The filament is being retracted mid print but this fails due to no cut or reversal of the extruder.

I believe both are caused by some kind of issue with filament detection. I’ve replaced the filament detection sensor since the video above, so its brand new again.

I have a THV8 board, with the janky wire bundle connector between boards. I have a suspicion that this connector is a bit flakey and the filament sense signal is getting lost at the board or connector/cable level.

If the printer loses this signal, it could cause #1 by the filament going through to the extruder but not being detected by the sensor (or rather, detected but that signal is not getting all the way back to the main processing component), and it could cause #2 by mid print the printer ‘detecting’ a loss of filament and trying to reload, but because the filament is still actually present, the logic is incorrect (no cut or retract) and causes the issues.

I wish the filament sensor had an led or something in it to indicate its triggered, that would help this debugging a bit if there was a discrepancy between the led and the screen ui for detection.

Aside from the toolhead filament sensor, the only sensor for the filament being loaded is, to my knowledge, the ams hub / buffer? Is there any other point where the filament is detected?

Does anyone have a pinout for this TH V8 board connector? is the signal for filament detection assigned to one of the pins? It seems to make sense that this is how it works based on the number of wires in the cable, and the components on each board front and back. I’ve googled but come up blank.

If I buy a V9 board, thats $90. and I will also have to replace my lidar as far as i can tell, which is another $90. So I’d like to be super confident before doing this.

I have a ticket open with support, but honestly since about 6 months after launch the support has gone downhill dramatically. I don’t expect any resolution there.