2nd Filament fails to reach extruder after bowden length calibration  

Re: 2nd Filament fails to reach extruder after bowden length calibration

I have had the same issue as well. I have attached an image of the measurements that I have. Looking forward to finding the solution to this issue.
Cheers Andrew

Hmm with gears in this configuration you should not have issues. The smallest diameter gear is in the first filament position. Did you do bowden length calibration? If yours doesnt work, there's no hope for the rest of us haha

Re: 2nd Filament fails to reach extruder after bowden length calibration

If the current design is so sensitive to gears precision, the idler bearings could be suspended by independent springs.

Re: 2nd Filament fails to reach extruder after bowden length calibration

Hi all,

I would like to ammend my earlier statement...

Adjusting the FINDA seemed to be the right way to improve reliability by a long shot.
My FINDA was adjusted according to the manual by insertig filament, screwing the FINDA in until the filament got caught, then releasing one quarter turn.

From my experiences and observations so far, this posed an enormous problem with PVA. Since that is quite stringy, it would leave a sometimes short and sometimes a looong string.

With the original setting, I had the impression that the FINDA would not detect the full diameter of the filament, but rather at the smallest thing.
So I moved the FINDA one turn out (equaling 1mm), which made it now trigger only when almost the full diameter is sliding past.
Looking at the ends of filaments that are created by the unloading, it appeared to me that the curveature on the full diameter is almost perfectly consistent with all kind of filaments. Only when the diameter of the cone gets smaller, the curveature changes and differ from filament to filament.
But when it comes to the bowden-tube, the distance from the FINDA to the extruder is only relevant on the full diameter of the filament, not when the first bits of filament arrive...

This has two consequences:
First, the FINDA must be adjusted differently then to what the manual says.
Second, the optical sensor at the hot-end can't be used as it can't measure diameter and might thus trigger before the full diameter has reached. So back to square one, when every tube must be manually calibrated and stored as a fixed number.
...but if it works that way - it's good enough for me... 🙂

Cheers,
Daniel

Re: 2nd Filament fails to reach extruder after bowden length calibration

Is there someone actively working on a firmware fix for this?
I couldn't find any relevant commits or PRs on https://github.com/prusa3d/MM-control-01 .
It shouldn't be to hard to implement the individual calibration for each gear, maybe I'll try that next weekend when I have the time.
I tried to make a backup of the firmware today, but I couldn't get avrdude to connect. Further investigation is needed. 🙂

Re: 2nd Filament fails to reach extruder after bowden length calibration

Talked to the support, really nice chaps. My hobbed gears are out of spec, I'll be getting a new set.

The nominal diameter for the gears is about 6.2.

I'll try installing the three 'good' ones in positions 1, 2, 3 for now to have at least a PMMU (P for partial) 😛

I also created https://github.com/prusa3d/Prusa-Firmware/issues/1194 with an idea how to automate the calibration.

Mine are 6.05, 5.80, 5.80, 6.15 and 6.20 (+/- by averaging several measurements). This exactly explains why I could't manage to print from positions 2 & 3. The support has changed their opinion. Now, that is perfectly within tolerances:

"the tolerances are quite within the expectations
we need to narrow down the issue and the supplier side
to have more accurate tollerances within smaller incongruencies"

The operation was successful, but the patient died.

Re: 2nd Filament fails to reach extruder after bowden length calibration

Hi all,

I would like to ammend my earlier statement...

Adjusting the FINDA seemed to be the right way to improve reliability by a long shot.
My FINDA was adjusted according to the manual by insertig filament, screwing the FINDA in until the filament got caught, then releasing one quarter turn.

From my experiences and observations so far, this posed an enormous problem with PVA. Since that is quite stringy, it would leave a sometimes short and sometimes a looong string.

With the original setting, I had the impression that the FINDA would not detect the full diameter of the filament, but rather at the smallest thing.
So I moved the FINDA one turn out (equaling 1mm), which made it now trigger only when almost the full diameter is sliding past.
Looking at the ends of filaments that are created by the unloading, it appeared to me that the curveature on the full diameter is almost perfectly consistent with all kind of filaments. Only when the diameter of the cone gets smaller, the curveature changes and differ from filament to filament.
But when it comes to the bowden-tube, the distance from the FINDA to the extruder is only relevant on the full diameter of the filament, not when the first bits of filament arrive...

This has two consequences:
First, the FINDA must be adjusted differently then to what the manual says.
Second, the optical sensor at the hot-end can't be used as it can't measure diameter and might thus trigger before the full diameter has reached. So back to square one, when every tube must be manually calibrated and stored as a fixed number.
...but if it works that way - it's good enough for me... 🙂

Cheers,
Daniel

Great observation, I was thinking of the same thing initially (whether the slightest string would trigger the FINDA or whether it should be the full diameter). I'll be lifting my FINDA higher too after I receive my replacement bearings.

For reference, my bearing diameters were: 5.95, 6.02, 6.35, 6.35, 6.35.

Re: 2nd Filament fails to reach extruder after bowden length calibration

Is there someone actively working on a firmware fix for this?
I couldn't find any relevant commits or PRs on https://github.com/prusa3d/MM-control-01 .

I stand corrected 🙂
https://github.com/prusa3d/MM-control-01/pull/28

Thank you Pavel, I can't wait to try it out!
What's the correct way to update the MMU firmware?
Do I update it directly via USB, or will the next version of the printer firmware update the MMU for me?

Re: 2nd Filament fails to reach extruder after bowden length calibration

What's the correct way to update the MMU firmware?
Do I update it directly via USB, or will the next version of the printer firmware update the MMU for me?

You should have a micro-USB cable to connect to the MMU and flash with Slic3r.

Peter

Re: 2nd Filament fails to reach extruder after bowden length calibration

I stand corrected 🙂
https://github.com/prusa3d/MM-control-01/pull/28

But they still haven't addressed the problems in load_filament_inPrinter() 🙁

Peter

Re: 2nd Filament fails to reach extruder after bowden length calibration

Okay, I could compile the current version from github without problems. Flashing with slic3r failed, but flashing directly from the Arduino IDE worked fine. I could calibrate the gears (only tried the first three though), and now I'm halfway through my first 3-color print and it looks very promising.

Re: 2nd Filament fails to reach extruder after bowden length calibration

Hey Peter

I'm having multiple print failures that I think are due to the difference in gear size on the filament feeder. I have arranged them so the smallest is on filament #1, but by the time it gets to my 3rd, 4th, and 5th filament when it loads it grinds a significant chunk off the filament as its ramming the filament into the bondtech gears that aren't spinning. I see that there is an update out there on github, but I have no idea how to compile my own file, and it looks like the prusa firmware isn't updated yet. Is there a way you could send me a compiled .hex file I could flash to my MMU. I'm really eager to see if individual feed lengths will solve my misprints, or if I need to continue looking for something else.

Thanks!
Tim

Re: 2nd Filament fails to reach extruder after bowden length calibration

I got the latest mmu firmware beta from git and tried compiling it in Arduino IDE 1.8.7 but I get this error:

"adc.c:26:2: error: 'DIDR2' undeclared (first use in this function)"

I don't see DIDR2 declared anywhere else in the sketch, so I'm confused.

Re: 2nd Filament fails to reach extruder after bowden length calibration

I haven't tried compiling the latest as yet. But the previous version required the "Board" to be set to "Leonardo".

Will have a look in an hour or so.

Peter

Re: 2nd Filament fails to reach extruder after bowden length calibration

Here's the MM-control compiled.

I have not tested this, so please use with caution.

Peter

Re: 2nd Filament fails to reach extruder after bowden length calibration

Here's the MM-control compiled.

I have not tested this, so please use with caution.

Peter

Thanks Peter!!

Re: 2nd Filament fails to reach extruder after bowden length calibration

Here's the MM-control compiled.

I have not tested this, so please use with caution.

Peter

Thank you very much Peter!

Re: 2nd Filament fails to reach extruder after bowden length calibration

https://github.com/prusa3d/MM-control-01/blob/master/README.md

Instructions on how to compile the source. "Leonardo" might work, but it isn't "right".

Building
Arduino
Recomended version is arduino 1.8.5.

Adding MMUv2 board
In Arduino IDE open File / Settings
Set Additional boards manager URL to:
https://raw.githubusercontent.com/prusa3d/Arduino_Boards/master/IDE_Board_Manager/package_prusa3d_index.json
Open Tools / Board: / Boards manager... Install Prusa Research AVR Boards by Prusa Research
which contains only one board:
Original Prusa i3 MK3 Multi Material 2.0

Select board Original Prusa i3 MK3 Multi Material 2.0

Build
click verify to build

Flashing
Windows
Arduino
click Upload

Slic3er
Hex file needs to be edited to be recognized as for MMUv2, to be specified later (in several years)

Avrdude
Board needs to be reset to bootloader. Bootloader has 5 seconds timeout and then returns to the application.

This can be accomplished manually by clicking reset button on MMU, or programmatically by opening and closing its virtual serial line at baudrate 1500.

Than flash it using following command, replace <virtual serial port> with CDC device created by MMU usually com<nr.> under Windows and /dev/ttyACM<nr.> under Linux. -b baud rate is don't care value, probably doesn't have to be specified at all, as there is no physical uart.

avrdude -v -p atmega32u4 -c avr109 -P <virtual serial port> -b 57600 -D -U flash:w:MM-control-01.ino.hex:i
Linux
Same as Windows, but there is known issue with ModemManager:

If you have the modemmanager installed, you either need to deinstall it, or blacklist the Prusa Research USB devices:

/etc/udev/rules.d/99-mm.rules

# Original Prusa i3 MK3 Multi Material 2.0 upgrade
ATTRS{idVendor}=="2c99", ATTRS{idProduct}=="0003", ENV{ID_MM_DEVICE_IGNORE}="1"
ATTRS{idVendor}=="2c99", ATTRS{idProduct}=="0004", ENV{ID_MM_DEVICE_IGNORE}="1"

$ sudo udevadm control --reload-rules
A request has been sent to Ubuntu, Debian and ModemManager to blacklist the whole Prusa Research VID space.

https://bugs.launchpad.net/ubuntu/+source/modemmanager/+bug/1781975

https://bugs.debian.org/cgi-bin/pkgreport.cgi?dist=unstable;package=modemmanager

and reported to https://lists.freedesktop.org/archives/modemmanager-devel/2018-July/006471.html

Building documentation
Run doxygen in MM-control-01 folder. Documentation is generated in Doc subfolder.

FYI - the link that is shortened is: https://raw.githubusercontent.com/prusa3d/Arduino_Boards/master/IDE_Board_Manager/package_prusa3d_index.json

Re: 2nd Filament fails to reach extruder after bowden length calibration

"Leonardo" might work, but it isn't "right".

There's a bit of a clue in MM-control-01.ino:

// MM-control-01
// HW arduino Leonardo
//
// peripherals:
//
// 3 x TMC2130
// pinout:
// pin apin port bit signal
// 0 0 C 0 CS0
// 0 0 C 0 CS1
// 0 0 C 0 CS2
// 0 0 C 0 STP0
// 0 0 C 0 STP1
// 0 0 C 0 STP2
//
// 16-bit shift register (2x 74595)
// pinout:
// pin apin port bit signal
// 0 0 C 0 CLK
// 0 0 C 0 DATA
// 0 0 C 0 LATCH

Please note line 2.

Peter

Re: 2nd Filament fails to reach extruder after bowden length calibration

That's new. Possibly cause of my whining on GitHub. The instructions I quoted appear to be in response as well.

After using your settings for getting the lower gears moving (Speeds of 19 and 22) I've effectively eliminated the grinding - but I also changed the amount it returns under filament settings:

Between that and your settings for filament my filament is coming out pretty reasonable.

But I'm having massive strings on cheap filament. It works but not well, a hang ~1/15 changes/

Re: 2nd Filament fails to reach extruder after bowden length calibration

Thank you! The compile instructions worked. I was able to adjust the bowden tube length for the first two filaments.

Question: Is there a button combination to 'save' the calibration? Or is it saved when you hit the middle button to retract the filament after setting the length?