MK3S/MMU2S not loading filament on print  

Page 1 / 3
  RSS
david.l65
(@david-l65)
Active Member

Reposting in MMU2S forum from MK3S forum

I just upgraded my MK3/MMU2 to MK3S/MMU2S. When I start a print, the filament does not load correctly unless I use "load filament to nozzle" first. If I don't do this, then the first layer does not print. I have only used single color printing since the upgrade, so I don't know what would happen with a MMU2S filament change.

Things I have considered:
-MK3S IR sensor is not working correctly. But I verified that the MK3S IR sensor works correctly (shows 0 when there is no filament and shows 1 when there is filament in the Bondtech gear).
-Filament is binding somewhere (reel to MMU2S, MMU2S to extruder) But the filament feeding works fine for a whole print if I load it first, it's just the initial loading that fails.
-Bondtech or MMU2S pinch rollers do not have enough pinch pressure. But it prints fine if filament is loaded to nozzle first.

This just seems like the firmware is not feeding enough filament after the MK3S IR sensor triggers.

Posted : 18/03/2019 1:00 am
PJR
 pjr
(@pjr)
Antient Member Moderator

Can you please confirm that you have installed the MMU2S extruder parts:

Not that the IR sensor is triggered by the lever on the Bondtech idler door and not with the ball and magnets (which should not be used).

Can you also confirm that you have uploaded the "S" firmwares for both printer and MMU.

Peter

Please note: I do not have any affiliation with Prusa Research. Any advices given are offered in good faith. It is your responsibility to ensure that by following my advice you do not suffer or cause injury, damage…...
Posted : 18/03/2019 9:59 am
david.l65
(@david-l65)
Active Member

Not that the IR sensor is triggered by the lever on the Bondtech idler door and not with the ball and magnets (which should not be used).

Yes, the IR sensor in the MK3S (in menu LCD menu - Support - Sensor) read 0 with no filament and 1 with filament loaded in the Bondtech gear. There is no ball and magnet in my upgrade...just the lever and IR sensor.


Can you also confirm that you have uploaded the "S" firmwares for both printer and MMU.

I upgraded both MK3S and MMU2S with the firmware in this location:
Original Prusa i3 MK3S/MK2.5S Multi Material 2S
http://prusa3dprinters.vshcdn.net/downloads/firmware/prusa3d_fw_3_6_0_MK3S_1_0_5_MMU2S.zip

However, the MMU firmware in that package is labeled prusa3d_fw_MMU2board_1_0_5.hex not prusa3d_fw_MMU2Sboard_1_0_5.hex as I would have expected. The firmware on the main Info Screen says Prusa i3 MK3S, but in LCD Panel-Support it says MMU2 connected, not MMU2S connected. Is the problem that the MMU2S firmware is not included in the package?

Posted : 18/03/2019 12:36 pm
david.l65
(@david-l65)
Active Member

Solved:

Performed the MMU2S factory reset and now it works correctly.

Posted : 18/03/2019 1:45 pm
christian.h7
(@christian-h7)
New Member

Same problem here. Even a reset did not help.
IR is 0 without filament, 1 with filament.
Reset the MMU2S to factory defaults. Did not help.
Reset the MMU2S to factory defaults and calibrated the filament length. Did not help.
Might be a bug in the MMU firmware (?) since the filament is been taken in a few centimeters but then the MMU2S won't continue to load it to the nozzle.

...
Posted : 04/04/2019 2:34 pm
beat.g and ewald.i liked
chris.b3
(@chris-b3)
Active Member

I'm having same problem. Resetting MMU2 didn't help. I confirmed that the IR sensor is not faulty. However, when it's in the "chimey" it isn't getting triggered by the idler lever. It appears that either the chimney is either slightly too high or the lever isn't long enough. I did manage to get it working if I used electrical tape to hold the IR sensor board down tighter....obviously not ideal.

...
Posted : 04/04/2019 6:43 pm
christian.h7
(@christian-h7)
New Member

The culprit indeed is the IR sensor. Try the self test, it probably will fail.

What happens is: It may show 0 when the extruder is not moving, when the extruder moves it may switch to 1 because of the vibrations.

There's very little tolerance here, I did not get my chimney to move further to the left of the extruder.

I fixed the issue for me using the following method (no guarantee that it works for you, please be careful and do small iterations):

  • go into the support menu, find the sensor info and select it (we'll need that the whole time)

  • unscrew the spring-screw of the extruder-door (side of the extruder)

  • unscrew the screw holding the door in place (back side of the extruder) but keep it close to you

  • use the file to remove a bit of the material (just a few micrometers) => green arrow

  • put the door back in and use the holding screw to fix (tighten it like it was before and be aware that the door can still move) the door again

  • apply pressure directly on the nut using your finger => blue arrow

  • use your finger and apply very gentle pressure on the lever (opposite of the green arrow, ATTENTION: the door should not move)

  • repeat the steps until your IR shows 0 even if you have followed the last step

  • use the selftest to check if the IR works even if the extruder is loading
  • repeat those steps until the self-tests pass and you can load your filament to the extruder using the menu.

    Again, attention: If you remove too much of the material, the IR sensor will not get a "1" again.
    In this case you can either move the chimney a bit to the right or reprint the part and start from scratch.

    Hope this helps you as well.

    BR,
    Christian

    ...
    Posted : 04/04/2019 8:26 pm
    Blizz
    (@steve)
    Estimable Member

    ^ So much this!

    I was testing the "load to nozzle" function and the MMU2S stopped feeding just after the FINDA was triggered, as if the bowden was 2cm or so long.
    I proceeded to do a factory reset, as I figured it might still have settings from before the upgrade.
    That didn't work either, but since I calibrated the IR sensor I was looking for other issues.
    Reading the above post and verifying this, it turned out that the IR sensor had a hair trigger indeed, often going to 1 because of the printer vibration (I went into the Sensor info while it was "printing air" and found the IR sensor constantly toggling between 0 and 1)

    So, the Support > Sensor Info here is of vital importance!

    I worked a little bit differently as described above but essentially did the same.
    I kept the tension screw etc as is, since I already tested & calibrated my MK2.5S directly, bypassing the MMU2 (so I was certain the extruder worked correctly.

    Without a filament in it, I unskrewed the chimney, taking off a small amount of material from the lever with an x-acto knife and putting the chimney back as "left as possible".This until the IR stopped indicating 1 (during calibration I got it to stay on 0, but no longer).

    I then started pushing filament in it until it arrived between the gears (and it pushed open the side door).
    Repeated to take of more material if it toggled too fast (as in: barely touching the gears) until I reliably got a "1" with filament and "0" without.
    Some shaking etc later (with and without filament) I was fairly confident, mounted everything again and sure enough, it worked.

    As mentioned above, be sure to only remove material in small amounts and test often. I think that in the end I didn't even remove 0.5mm, so it's not much.

    Thanks for pointing me in the right direction!

    MK3S / MK2.5S-MMU2S / Ultimaker 2+...
    Posted : 05/04/2019 1:23 pm
    mark.k22 and sneaks liked
    mark.m30
    (@mark-m30)
    Eminent Member

    Ok, I'm having this same issue. I have had to load the filament to nozzle to get the print started, but now it started printing the wrong color and when the color switch came it performed beautifully with the wrong color. Basically, the colors have switched. I printed the provided sheep g-code and that printed just fine, so I'm sure all my sensors are working fine. I compared the sheep code to the code I just tried to print and they were identical in the beginning.

    Posted : 17/04/2019 4:08 am
    mark.m30
    (@mark-m30)
    Eminent Member

    I just tried the Sheep again, same code, but not feeding the first filament to the extruder. The IR sensor is fine as far as I can tell.

    Posted : 17/04/2019 2:11 pm
    mark.m30
    (@mark-m30)
    Eminent Member

    So after looking closer at the issue I have determined that it is indeed the IR sensor getting triggered before the print starts. This is happening because when the extruder is over to the left, the bowden tube is pulling the chimney over enough to trigger the sensor. It seems there is not enough landing for the chimney, just the top of the orange block and it is rocking there. This is rather poor mechanical design of a critical part. Trying to solve this I double stick taped two match sticks to either side of the orange insert to give a broader base for the chimney. So far this seems to work.

    Posted : 17/04/2019 4:15 pm
    daniel.k45
    (@daniel-k45)
    Eminent Member

    My problem is that it always shows 0 even with filament loaded. Anyone have this issue? I try to glue some extra plastic on the little arm. But it works with no motion but when you wiggle the chimney a little bit it jumps to 0 and 1 all the time. 

    Posted : 24/04/2019 4:35 pm
    hamid.m
    (@hamid-m)
    New Member

    Same problem here MMU2S not loading any of 5 filaments everything is in order including IR and FINDA.

    This post was modified 1 year ago 2 times by hamid.m
    Posted : 03/05/2019 4:45 am
    daniel.k45
    (@daniel-k45)
    Eminent Member

    I gave up. the MMU is so buggy. It works for one print then it stops and you have to load all the filaments all over again. Then it loads unloads in a loop. The motors seem to be on all the time and get very hot even when the printer is not printing. I removed it and now printing with single again. 

    Posted : 03/05/2019 12:27 pm
    sneaks
    (@sneaks)
    Active Member

    Christian & Steve - Thanks so much!  I did the MK3S upgrade today and have been banging my head against the wall for the past 6+ hours.  I was lucky enough to stumble onto your post and sure enough, I just had to trim down some plastic.  Mine was quite a bit, I started sanding then moved to a exacto knife.  Regardless, it worked and my printer is printing!

    I am watching the Sensor"IR" while it is printing and it is going back and fourth between 0 and 1 at certain points.  I think I am going to print some more parts to test.

    ...
    Posted : 01/06/2019 11:38 pm
    sneaks
    (@sneaks)
    Active Member

    Turns out on my printer, the issue wasn't the chimney needing trimming.  The issue was the idler door wasn't closing far enough to move the chimney out of the sensor.  I did a bunch of trimming and sanding on the area around the gear to solve the issue.

    I made a modified "extruder_idler_mmu2s1.stl" that fixed the issue for me:  https://www.prusaprinters.org/prints/3417-mk3smmu2s-modified-extruder-idler-mmu2s1-solves-ir

    This post was modified 1 year ago by sneaks
    ...
    Posted : 02/06/2019 2:32 pm
    Antimix
    (@antimix)
    Estimable Member

    I am testing the MMU2S since a couple of days. Of course there are issues. The concept is OK, but the technical realization still need several improvements. What I found:

    - I had to carefully slide enlarge (in horizontal) the holes IR sensor cover to be able to correctly place it on the IR board.

    - I had to cut the soldered pins of the IR board, making them around 0.5mm so that they do not hit the plastic  cover, and it can be closed and placed correctly.

    - Also my MMU2S gears do not pinch the filament on the rollers with the correct pressure, and the filament slip away (stay still)
    Considerations: I have tried to tighten the pressure screws with springs on MMU but it did not solve the issue, since the filament is smooth and the bearings are also smooth, so the gears from one side is not enough to move the filament. PRUSA should consider to have a MMU central body  with double gears instead of bearings. That could give more grip.

    - The filament often is not loaded to the nozzle, but is stays in the middle of the bowden tube.
    It is clear that it happens because the filament is shifting on the MMU gears. When I see it, I have to help it pushing it manually till it reach the extruder gear and it is loaded. The MMU2S has no way to tell that the filament is shifting.

    - The bowden tubes diameters ARE TOO SMALL. PRUSA should find bowdens with a  larger internal diameter.
    The filaments pass through the tube with almost no space, and just flexing the bowdens  causes the filament to be almost blocked by the friction. I can hear how much force I should apply to move the filament, and that is no good for the gears that will surely slip. I removed the long bowden and the buffer, and inserted just five 10cm bowdens, so that the filaments from the spools have no impediments (will be changed in future, but now I need to eliminate disturbing factors). The spool are higher on a bar, so there is free space for the filament for now.
    However also with this, the filaments do not move well in the bowden between the MMU and the extruder, and it not load well and stuck in it: why ? Because the pinch of the gear on the extruder, and the gear of the MMU, create after several retractions on the filament a notched serration, that deforms the shape of the filament, and just this is enough to not be able to pass through the tight tube diameter without a  strong force (not supplied by the MMU2S gears, since it shifts on the smooth  bearings). This happens usually after the first print, when a filament has been retracted to the MMU2S.

    - The BLOODY Blade cutter. It slip away after some cuts. The way it is hold in place have to be redesigned.
    Due to the fact that the piece is 3D printed, the tolerance is not reliable. Mine, after some cuts, slipped away, and I realized that is almost impossible to reinstall it in that position, at 45° degree, in lower position, where the gravity drops the blade on the holder down that is completely flat while the shape that hold it is up , and become impossible to install, unless half dismount several parts (e.g.  bars, motor.... 30min work at least).
    I expect also that the blade metal will wear out the PET borders make it impossible to install.
    PRUSA have to decide if create a custom cutted blade with holes that can be screwed (and you will be able to get only from them) or chose a different commonly found blade and try to adapt the holder part.

    - The filament often stuck (I hear it hitting the plastic) trying to enter into the hole of the MMU2 FINDA section.
    Enlarge the in hole of the MMU2 FINDA section, making it with a larger diameter, and conical to better address the filament in the center.

    It is all for now.
    Regards

    Posted : 02/06/2019 9:06 pm
    dan.e4 liked
    admin
    (@admin)
    New Member
    Posted by: Antimix

    I am testing the MMU2S since a couple of days. Of course there are issues. The concept is OK, but the technical realization still need several improvements. What I found:

    - I had to carefully slide enlarge (in horizontal) the holes IR sensor cover to be able to correctly place it on the IR board.

    - I had to cut the soldered pins of the IR board, making them around 0.5mm so that they do not hit the plastic  cover, and it can be closed and placed correctly.

    - Also my MMU2S gears do not pinch the filament on the rollers with the correct pressure, and the filament slip away (stay still)
    Considerations: I have tried to tighten the pressure screws with springs on MMU but it did not solve the issue, since the filament is smooth and the bearings are also smooth, so the gears from one side is not enough to move the filament. PRUSA should consider to have a MMU central body  with double gears instead of bearings. That could give more grip.

    - The filament often is not loaded to the nozzle, but is stays in the middle of the bowden tube.
    It is clear that it happens because the filament is shifting on the MMU gears. When I see it, I have to help it pushing it manually till it reach the extruder gear and it is loaded. The MMU2S has no way to tell that the filament is shifting.

    - The bowden tubes diameters ARE TOO SMALL. PRUSA should find bowdens with a  larger internal diameter.
    The filaments pass through the tube with almost no space, and just flexing the bowdens  causes the filament to be almost blocked by the friction. I can hear how much force I should apply to move the filament, and that is no good for the gears that will surely slip. I removed the long bowden and the buffer, and inserted just five 10cm bowdens, so that the filaments from the spools have no impediments (will be changed in future, but now I need to eliminate disturbing factors). The spool are higher on a bar, so there is free space for the filament for now.
    However also with this, the filaments do not move well in the bowden between the MMU and the extruder, and it not load well and stuck in it: why ? Because the pinch of the gear on the extruder, and the gear of the MMU, create after several retractions on the filament a notched serration, that deforms the shape of the filament, and just this is enough to not be able to pass through the tight tube diameter without a  strong force (not supplied by the MMU2S gears, since it shifts on the smooth  bearings). This happens usually after the first print, when a filament has been retracted to the MMU2S.

    - The BLOODY Blade cutter. It slip away after some cuts. The way it is hold in place have to be redesigned.
    Due to the fact that the piece is 3D printed, the tolerance is not reliable. Mine, after some cuts, slipped away, and I realized that is almost impossible to reinstall it in that position, at 45° degree, in lower position, where the gravity drops the blade on the holder down that is completely flat while the shape that hold it is up , and become impossible to install, unless half dismount several parts (e.g.  bars, motor.... 30min work at least).
    I expect also that the blade metal will wear out the PET borders make it impossible to install.
    PRUSA have to decide if create a custom cutted blade with holes that can be screwed (and you will be able to get only from them) or chose a different commonly found blade and try to adapt the holder part.

    - The filament often stuck (I hear it hitting the plastic) trying to enter into the hole of the MMU2 FINDA section.
    Enlarge the in hole of the MMU2 FINDA section, making it with a larger diameter, and conical to better address the filament in the center.

    It is all for now.
    Regards

    I just finished my MMU2S upgrade on MK3S and I 100% agree with you. It's not reliable I can't use it for long autonomous print...

    There is two CRITICAL point :

    - The bowden tubes diameters ARE TOO SMALL

    - Filament fail trying to enter into the hole of the MMU2 FINDA section -> need a large conical hole + PTFE chamfer on TWO SIDE !!

    (sorry for my english)

    Pascal

     

    Posted : 10/06/2019 9:15 am
    mkopack
    (@mkopack)
    Eminent Member

    Same problem... Did the Mk3+MMU2 upgrade to the S...

    Before the upgrade I could reliably get filament to load down to the extruder in single color mode and that's how I printed all the time without issues.

    Now I can't get a single damn filament to even LOAD properly!!!

    IR sensor seems to work fine - getting 0 when no filament in, 1 with it in.

    When I try to load anything, it shoves it down the orange PTFE tube to the extruder, but it doesn't seem to get taken up by the bondtech gears, it then immediately starts backing back out up to the MMU2. The Bondtech gears never grab and pull in the filament. I've tried cutting the tips at different angles, etc. Doesn't seem to matter.

    I'm beyond frustrated and ready to just take things back the Mk3 (no S). The whole S upgrade just seems really poor. I don't get why the hell they went with this IR sensor for MMU use, but use the PINDA+magnet for non-MMU2S use? As things currently stand I don't see how we would EVER be able to print flexible with this setup (with the plain Mk3 I had to always set the bondtechs super loose, which you absolutely CAN'T do with this setup or it won't trigger the IR sensor... ) 

    They really should have made the IR sensor something that you could adjust using a bolt through a nut that you could adjust in/out with fine adjustment that would not be tied to the bondtech pressure.

    I LOVED my Mk3... I even liked it with the MMU2 even though I wasn't doing multi-color. This setup I want to throw through a window and smash into 1000 pieces!

    Posted : 29/06/2019 3:32 am
    sneaks
    (@sneaks)
    Active Member

    It sounds like you may need to run the filament calibration.  This finely tunes the position where the filament stops by the bondtech gears and is crucial for proper loading: FILAMENT CALIBRATION GUIDE

    After I did the 'S' upgrade I had to run through this to get the filaments to load.

    ...
    Posted : 29/06/2019 1:22 pm
    Page 1 / 3
    Share:

    Please Login or Register