Prusa-firmware: [BUG] MMU2S filament load and change fails 90% of the time

[linzwatt]

Printer type - MK3S
Printer firmware version- 3.8.1

MMU Upgrade - MMU2S
MMU upgrade firmware version - 1.0.6

SD card or USB/Octoprint - SD CARD

Describe the bug
When starting a print, the first filament load fails most of the time, if the first filament load works, any filament changes during the print are most likely to fail too. The MMU2S pushes the filament all the way into the extruder, activates the sensor (confirmed by looking at sensor debug menu), filament goes back and forth a little then is drawn all the way back into the MMU2S. This repeats 3 times before a filament load failure is reported on the LCD.

However, when I do a "Load to nozzle" before starting a single filament print, this works 100% of the time.
If you start a print straight after a "Load to nozzle", selecting the same filament you loaded to nozzle, the print starts straight away without loading again.

I imagine the "Load to nozzle" uses a slightly different procedure to the loading process at start of print and filament change process during print. Can any developers confirm whether the process is different? Or perhaps the timing is different? It could be that the sampling window for extruder filament sensor is not long enough during a load at start of print.

To Reproduce

1. Start a print with no filament loaded
2. Printer will try to load the selected filament - likely to fail
3. Stop the print
4. Do a "Load to nozzle" command
5. Printer will load the selected filament - likely to work 1st time
6. Start a print
7. Printer starts the print with the pre-loaded filament, no issues
8. Subsequent filament changes during the print are still likely to fail

Expected behavior
Printer should load filament successfully 100% of the time

[sudo-gedit]

I can confirm the mistake too.

[dikodahan]

Not sure if this is tied to the same issue listed above, but I am experiencing a similar issue at the end of a print.
This has just started after the upgrade to MK3S 3.9.1 (MMU2S at 1.0.6). The entire print will complete successfully and then in the unload at the end I am getting an error message saying MMU2 needs attention and I have to manually pull the filament out.
This happened on a single filament print, so I assume it is the same bug type.

The bigger issue is that the MMU is "stuck" after. The only way to get it back to normal operation is a complete disconnect and reconnect from power. A simple button reset does not work!

[dikodahan]

Any update on this one??
This is a seriously irritating behavior!

[Syl20-94]

Hello, I've got tha same bahaviour, with FW 3.9.1 and MMU 1.0.6, each action on the MMU cause strange behave and all leds of mmu unit blink fast (orange+green) then MMU reboot a couple of time before going back on the normal way - usually I need to reset manually the MMU to go on.

[dikodahan]

I can't do anything now with the MMU... any type of action causes all LEDs to blink orange.
Prusa team, help here please...??

[linzwatt]

I think the most recent comments are separate issues, although they all relate to very unreliable MMU firmware
would be good to get some comments from the main contributors to the MMU firmware

@wavexx @josefprusa @PavelSindler @DRracer @mkbel ?

[ANTALIFE]

Sounds like you have bad tips, bad as in the diameter is irregular. Not long ago I made some changes to 3.9.0 to get around this

https://www.antalife.com/2020/08/project-mmu2s-filament-diameter.html

[sudo-gedit]

Is there no interest in fixing the problem?

[Syl20-94]

@ANTALIFE , your point is interesting, but I'm not sure about it. in the process, filament is loading and detected by both PINDA and FINDA sensor (as motor MMU stop loading) then it goes back and do a reload twice with the same process (is it a kind of loop ??) and goes in failure (all led blinking) which need to reset the MMU unit.
after the reset and without touching the filament (neither cutting/replacing) the MMU is loading normally (but sometime I need to reset a couple of time again)

So Yes I'm interested in fixing this problem but I'm not the developper so I can not goes deeper in fixing the only thing I can do is living with this workaround which is irritating me on each printing

[ANTALIFE]

@Syl20-94 Show us how your tips look when you see a load error/loop

[mkbel]

@sudo-gedit Can you supply log of the problem? How to obtain log: https://github.com/prusa3d/Prusa-Firmware/issues/1689#issuecomment-680850966

[dikodahan]

Guys, if it was a bad tips issue then the MMU2 would have had issues prior to 3.9.1 firmware. I've had the MMU2 (and then upgraded to MMU2S) for 2 years now. no issue at all, up until I upgraded to 3.9.1 on the printer, all of a sudden my tips are the issue??

In addition to the above, I have also, as an additional precautionary measure, took all the filaments out from it, clipped them, and put them back in. still the same problem.

[mkbel]

@dikodahan Can you supply log of the problem? How to obtain log: https://github.com/prusa3d/Prusa-Firmware/issues/1689#issuecomment-680850966

[Syl20-94]

@ANTALIFE indeed I can say it's a tip, here is my workflow.
1- Prepare print through PrusaSlicer 2.2.0, my printer is a mk3S with FW 3.9.1
2- Send Gcode to octoprint 0.17 from Prusaslicer via network
3- On octoprint, launch the print and select filament via MMU2S Select Filament plugin
From this point printer and octoprint work in automatic
4- Octoprint set heating for extruder and bed, in the meantime,
5- Filament is loading, MMU unit goes to a start-up sequence and move from left to right and move to correct filament position, start to load the filament to the extruder.
6- When filament arrive to the extruder - the lead screw is moving as filament is loaded by the extruder motor- MMU motor stop
7- After a short time (1 or 2 sec) filament is unloaded from the extruder back to MMU unit
8- Then it redo step 5 to 7 two time
9- after 3 loops, filament is unload and MMU unit start to blink for all orange + green LEDs
10- I catch a screw pin and press reset button on the MMU unit
11- MMU is rebooting and redo from step 5.
then it can redo either load/unload sequence til I need to reset MMU unit or it load correctly the filament
when filament is loaded, print sequence is going on without problem (currently, I print in 1 color mode so without filament change during printing).

so I dont touch filament during the whole process.
feel free to tell me if you need logs and how to catch it

[iplayfast]

I thought I was going mad. I'm glad to see that this is a bug that is affecting other people too.

[mkbel]

@iplayfast I would be really happy to fix this problem. It would by very helpful if someone who is able to reproduce the problem would do so when having printer connected via USB cable to some computer and logging debug information present on that USB CDC communication channel. How to do that is described here: https://github.com/prusa3d/Prusa-Firmware/issues/1689#issuecomment-680850966

[Syl20-94]

update from my side, I find out that the spring on the idler-body which is on the motor side was missing, I put it back and after 4-5 print, loading from scratch seems to be OK and it's goes back to normal working. I'm not explaining the reason but I recommand to check if the idler-body is not too tightened screwed (or too loosed)

[linzwatt]

@mkbel I will supply some logging info in the next couple of days, thanks for your interest and the info about how to collect logs :)

[chriswal]

if all leds are blinking on the MMU2 there is a problem with the motor drivers. if the 5v drops the drivers lose their configuration. this problem often comes from a loose plug. reseat your 5v plug on the mmu2 and the einsy board. or the drivers are overheated. touch your board of the mmu2 with your finger (but be carefull it could be very hot) if you burn it ;-) you know its overheated. (mine got 80degrees celsius in idle).

[sudo-gedit]

are there any new findings?

[mkbel]

@sudo-gedit Still no one produced log. I am main maintainer of MMUv2 firmware. I would be very happy to fix this bug if present. But I desperately need log to fix the problem.

[dikodahan]

I will produce an output log by the weekend for you on this one.