Mk3s/MMU2s filament load interrupted, filament reverses
 

Mk3s/MMU2s filament load interrupted, filament reverses  

  RSS
John Svensk
(@john-svensk)
Eminent Member

I have a problem with my Mk3S/MMU2S where filament is pulled/fed
to the IR sensor and possibly a bit further, but then the filament
is reversed and pushed back out again.

Interestingly, Finda and IR sensor calibration is ok!
I have spent six hours on this problem with Prusa Support, and
I haven't come a millimeter closer to a solution...
I have written a very condensed version of my chat transcripts below,
hoping that one of you big-brained people will find what Support
didn't. 😊

// ------------------------------------
// chat session 1

John: Hello. I'm having problems loading filament.
If I for instance use "load to nozzle", the extruder stepper starts,
and the MMU feeds filament to the extruder. Ok, so far so good.
But then the extruder stepper stops for a split second, and then
continues for 2,5 rotations. Then the stepper reverses direction,
feeding the filament out again! Then the stepper stops, the MMU
unloads the filament, and the extruder stepper returns to rotating
the normal direction. This happens twice, and then I get a
mmu load failed error. (...) Also, I have checked the IR sensor's
output with an oscilloscope, and I can't see any "jumping" of the
sensor's value...

(( comment: I had the time scale set to 200ms per square,
so I guess any transients of 20ms or shorter wouldn't show,
but I had triggers set on both rising and falling edges
of the signal, and they only reacted to the big obvious steps. ))

(...)

Support: Ok but the filaments do reach all the way to the extruder
and come back out again?

John: I'm not sure; they reach the IR sensor, I see the tension
screw move inwards just before the stepper halts

Support: Ok it's definitely going to have to do with the IR sensor
that's going to need some calibration.

John: already done that, but will do it again

Support: Great thanks. The problem is definitely in that area.
Have you updated the firmware to the latest versions too by chance?

John: yes

// chat session

2 Support: have you done the ir sensor calibration before,
and how did it behave?

J: Yes, but it was hard to find a good spot, so I tried removing
two layers from the arm on the idler door, and then I found a
good spot

(( comment: I had problems with feeding filament or the Allen key
through the IR sensor at times, but support person just
instructed me to disable the steppers, freeing the stepper that
was holding still, don't know if this is relevant. ))

Support: okay we will reflash the printer now please disconnect
the mmu from the printer do a factory reset on the printer and
flash the latest firmware. (...) then we need to do the
ir + finda calibration after the new firmware

// chat session 3

Support: so the FINDA and IR sensor are calibrated?

J: yes, but I'm still getting MMU load errors..

Support: can you describe what happens?

(( comment: After 50 minutes I was really struggling to be polite,
so I ended the chat session. ))
------------------------------------

After what was in total six hours (including heroic fights against firmware that "swallowed" menu items that were relevant for the debugging, i.e. "disable steppers") the error remains. So now I turn to you!

 

Best Answer by PJR:

The load process with firmware 3.7 is as follows:

MMU loads filament until IR sensor is triggered
Bondtech loads filament 50mm
Bondtech unloads filament 45mm
Printer checks IR sensor to ensure filament is still loaded

So, you need to ensure that the IR sensor is properly calibrated and that the filament paths are as free as possible.

Personally, I am still running firmware 3.6 which does not perform this nonsense.

 

Peter

This topic was modified 2 years ago by John Svensk
Posted : 28/05/2019 10:33 am
adam.sharp
(@adam-sharp)
New Member

I'm having the exact same problem, my next step was going to be to contact support.  Starting to pull my hair out trying to figure out why it's doing what it is...

Posted : 28/05/2019 12:10 pm
John Svensk
(@john-svensk)
Eminent Member

By all means, contact support! If  they hear that I'm not the only one, perhaps someone will put some energy into investigating... Have you looked at the issue lists on Github? That's my next step..

Posted : 28/05/2019 4:11 pm
gnat
 gnat
(@gnat)
Noble Member

I think you were on the right track with looking at the idler door. Two key points match what I experienced:

1) You had to trim the IR flag to get what appears to be a good reading.

2) Filament is very difficult to manually insert into the extruder.

Another key point for me was that even with the IR sensor apparently calibrated, I still could not get through a print without the MMU attached (disconnected from the main board) and IR sensor turned on. It would print a little bit and then tell me it ran out of filament when it clearly did not.

I ran off https://www.thingiverse.com/thing:3572252 and everything has been fine since. As the only difference in the model is the IR flag I can only assume that the difference is in the print quality, but in addition to allowing me to dial in the flag better the whole door fit better/looser and I can now manually insert filament just like I could before the S upgrade.

The suggestion I keep promoting is to disconnect the MMU entirely and get the printer working and dialed in as a plain MK3S including being able to get through an entire print (for me it was a simple 20x20x20mm cube) with the IR sensor enabled and no false run out events. Once the base printer is working, then add the MMU into the mix. Trying to troubleshoot when both are in question just complicates things.

MMU tips and troubleshooting
Stop SL1 tank leaks with a simple o-ring...
Posted : 28/05/2019 5:29 pm
PJR
 PJR
(@pjr)
Moderator

The load process with firmware 3.7 is as follows:

MMU loads filament until IR sensor is triggered
Bondtech loads filament 50mm
Bondtech unloads filament 45mm
Printer checks IR sensor to ensure filament is still loaded

So, you need to ensure that the IR sensor is properly calibrated and that the filament paths are as free as possible.

Personally, I am still running firmware 3.6 which does not perform this nonsense.

 

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 : 29/05/2019 9:37 am
John Svensk
(@john-svensk)
Eminent Member
Posted by: PJR

Bondtech loads filament 50mm
Bondtech unloads filament 45mm
Printer checks IR sensor to ensure filament is still loaded

So THAT is why it is so damned sensitive!

And to think that I have spent a total of six hours of (by some support staff) being treated like an idiot, when ”it’s not a bug, it’s a feature”... I’m just on my way in to a Crossfit workout, so I’ll work off most of my aggressions, then I’ll write a letter to mr. Prusa about the need for a second line support... 😤 🤯 

Posted : 29/05/2019 10:58 am
gnat
 gnat
(@gnat)
Noble Member

Yeah I think they really need a step between the S upgrade and adding the MMU to get people to stop and start printing as a plain MK3S to find and fix issues with the IR sensor. Once you get it working it seems to do it's reduced job great, but getting it there is a fiddly PITA. And until the IR sensor is playing nice trying to get the MMU working is just pointless frustration. 

MMU tips and troubleshooting
Stop SL1 tank leaks with a simple o-ring...
Posted : 29/05/2019 12:10 pm
John Svensk
(@john-svensk)
Eminent Member

@pjr 
After a bit of searching Github, I  found this:

github.com: Prusa-Firmware issue #1689

But the source code seems to point at a measly two millimetres of difference between what's pulled in and pushed out?

Posted : 29/05/2019 3:10 pm
John Svensk
(@john-svensk)
Eminent Member

Continued:
But never mind the millimetres, downgrading to 3.6.0 seems to have done the trick! 😍

Posted : 30/05/2019 7:57 am
PJR
 PJR
(@pjr)
Moderator
Posted by: John Svensk

Continued:
But never mind the millimetres, downgrading to 3.6.0 seems to have done the trick! 😍

🙂

 

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 : 30/05/2019 1:00 pm
Casey
(@casey-2)
Estimable Member

I found out that if you turn on Stealth mode, this sort of works a bit better on 3.7.1. I think it's picking up on the tension on the extruder or something like that?... Crash detection?

I found that interesting, because... I saw other posts saying that the extruder gets REALLY hot after the upgrade.

IR sensor definitely not the problem here. I got into the sensor menu to watch it during the process. It triggers/stays stable, and de-triggers when it should.

Eh. I gave up on support after I was laughed at and mocked, when I tried to give them a video of the thing... I guess they think it's 'funny' that I struggled with this for weeks, based on the nonpology I was sent.

Posted : 05/06/2019 1:43 am
John Svensk
(@john-svensk)
Eminent Member
Posted by: casey.b6

Eh. I gave up on support after I was laughed at and mocked, when I tried to give them a video of the thing... I guess they think it's 'funny' that I struggled with this for weeks, based on the nonpology I was sent.

Whaat? Laughed at? To be honest I don't believe you. I have spent hours trying to get my MMU2 in working order, but I have never been laughed at... Usually, support appreciates videos.

Posted : 05/06/2019 8:59 am
Casey
(@casey-2)
Estimable Member

Give me your email address and I'll forward the chat log if you don't believe me.

Posted : 05/06/2019 9:55 pm
Casey
(@casey-2)
Estimable Member

Found something out, unrelated. The extruder, toward the bottom, seems to be remelting (the PETG). Maybe this thing should be made of ABS or PC... or a little more distance/beef to it at the bottom... Similar to the warping that others have complained about earlier on the PINDA holder.

Posted : 05/06/2019 9:57 pm
Print McPrinty
(@print-mcprinty)
Active Member

Anyone still having these issues? I have been struggling with this for a while now. I'm about to go back to Firmware 3.6 as suggested above. 

...
Posted : 11/10/2019 3:10 pm
Dave Avery
(@dave-avery)
Honorable Member

the problem is a misaligned ir sensor every time. the MMU extruder door assy is VERY sensitive to setup. either moving the IR board around a it or modifying the flag on the door arm is the usual answer. the setup needs to ne 100% reliable ( always off if no filament in the gears and always on if there is filament). if the adjustment is a little off it can trigger on slight changes in filament diameter or gear eccentricity

Posted : 14/10/2019 5:23 pm
richard.l liked
shilbert
(@shilbert)
Active Member

@david-a66

If your really want to know what is going on:

https://www.prusaprinters.org/prints/2139-mk3smmu2s-filament-indicator-cover

Posted : 20/10/2019 6:22 am
Dave Avery
(@dave-avery)
Honorable Member

i've had the LED mod on my extruder for months ( but then I am using a BNBSX extruder on my MMU2S that has an easy adjustment screw for sensor cal)

Posted : 21/10/2019 7:39 pm
ACE liked
Print McPrinty
(@print-mcprinty)
Active Member

Turns out that replacing the PFTE tube in the extruded made a huge difference for me. Originally I had thought the newer tubes ID was too small. Having the tighter tube causes the tips to be much better formed. I also do two cooling moves and changed the retracting distance to -18.

...
Posted : 24/11/2019 1:49 pm
Share:

Please Login or Register