Notifications
Clear all

i3 mk2/s MM 1.0 - random layers not extruding  

  RSS
FaultyLine
(@faultyline)
Trusted Member
i3 mk2/s MM 1.0 - random layers not extruding

Hi everyone!

This is my first post, I'll try to include all relevant information - if I'm missing something please ask and I'll do my best to provide the desired info 😀

I'm having an intermittent issue where every 4th (or so) print fails due to any of the 4 extruder stepper motors failing to progress the movement of filament through the hotend. The failure is usually in the form of a partially completed print, and more recently, numerous layers missing in the middle of a print.

Before going too much further, here are the versions I'm running:
i3 MK2/S printer, with installed MMU 1.0 upgrade.
Printer firmware is 3.1.0.
Rambo13a motherboard.
E3Dv6full hotend.
Slic3r version 1.39.2+win64
Octopi version 1.3.8 running on RaspberryPi 3 model B v1.2

Before anyone asks - I have indeed restated every cable on the whole printer 😎

I've had failures from gcode generated using Slic3r printer selection mode of "Original Prusa i3 MK2 MM Single Mode" and "Original Prusa i3 MK2 MultiMaterial".

Here is one of the first failures from "Single Mode" using extruder 1:

Here is one of the next failures from "Single Mode" using extruder 4:

Here is a recent failure from the "MultiMaterial" mode using extruders 1-4. Every color had the band of missing material at the same layer height. Additionally, the purge tower had the very same artifact.

Every time a print has failed, I don't do anything with the printer at all and immediately go to Settings -> Move Axis -> Extruder, and command it to move. Every single time I've done this, the extruder moves, even after the print has failed. If I immediately restart the same print, using the same exact gcode, without rebooting the printer, it's likely the print will succeed. This confirms, to me, that the GCODE is OK.

I've verified the bowden gears are clean and clear of any debris, that filament easily moves through them, and that the spring tension is set via multi material manual specified value of 4-5mm from washer to bowden "extruder-idlers".

At this point, I'm convinced this is not a mechanical issue, but rather an electronics issue of some kind. I've based this off of the 3rd picture above, from my multi material failure. I used 4 different colors of PLA, and the print failed at the same layer for each of the 4 extruders. If this was indeed a mechanical issue, I would expect to see a single color failure, not all 4, in the same set of layers. I was able to very easily split the failed print down the missing layer line and saw some interesting stuff going on...

Here is the outside of the shell that I broke off at the missing layer line:

And did the same with the internal gears:

I can't explain, with confidence, as to what has caused the grey PLA to even be on the yellow and blue gears. My only guess is that it has to do with the same extruder problem I'm complaining about and is probably related to something causing the incorrect extruder to extrude material in the wrong part. Notice that it's oddly entirely missing on the aqua colored gears... ❓

I did some research and testing on how Prusa has made the RAMBo-mini achieve control of 4 separate extruder stepper motors, regardless of only having a single extruder output. The key to this problem is the "MK42 Stepper SW" board. In short, the MK42 Stepper Switch board is commanded via the Rambo-mini to connect, via relays, the desired extruder directly to the only extruder output on the RAMBo-mini board. Additionally, the MK42 Stepper Switch board *appears* to always have Extruder 1 connected to the RAMBo-mini in a default "resting" state, unless Extruder 2, 3, or 4 are required; at which point a relay(s) are activated. See pictures below:

Printer powered on:

Extruder 1 selected:

Extruder 2 selected:

Extruder 3 selected:

Extruder 4 selected:

You can see that the powered on state and extruder 1 selected state appear to be identical, according to the LED's on the MK42 Stepper Switch board. I am making an assumption that this board is doing "nothing" to connect extruder 1 to the RAMBO-mini board and that this is the default "resting" state of the MK42 Stepper Switch board.

Why is this signifigant? If my above assumption is correct, it means that Extruder 1 is connected to the Rambo-mini directly, without any relay switching being implimented. I've already had a failure on Extruder 1. I've also had a failure on Extruder 4, and recently a failure of all 4 extruders in the same layers of the gear bearing. All of this points me to look at the RAMBo-mini board, failing to properly drive the extruders. Given all 4 extruders failed at the same layers further supports this theory.

To troubleshoot this I have left the MK42 Stepper Switch control/selection cable connected to the RAMBo-mini board, but have taken Extruder 1 and removed it from MK42 Stepper Switch board, in effect bypassing it, and plugged it directly into the RAMBo-mini extruder output. This eliminates the MK42 Stepper Switch from affecting the stepper control signals, but nothing else.

If this next test print fails, it confirms, to me, that the RAMBo-mini has probably failed in some obscure way. If it does not fail - then something might be wrong with the MK42 Stepper Switch board.

Other variables that should be considered... (and probably tested)
Octopi - while it's been working for many months, just like the multi-material upgrade, I have submitted every print through this device. I suspect if I continue to see failures with my bypassed MK42 Stepper Switch board, I should put the GCODE on the SD card and re-run all of my tests to remove the Octopi variable from the equation.

After this giant wall of text and pictures - my questions are as follows:

Is there a way to confirm that the RAMBo-mini board is *actually* sending stepper signals through it's internal processing and that the real signals are getting generated and sent to the steppers? Maybe there is a system log somewhere I can acquire off the board after a failed print? I have a 4 channel 100 MHz oscilloscope if probing the physical connection if that's deemed necessary.

Is my current troubleshooting path seem correct?

Am I missing something?

Has anyone else had this problem, in any flavor of a MK2/S printer?

Are there any suggestions on what I can do as a next step?

Have I made a terrible assumption about how any of this works?

Thank you in advance for any advice, suggestions, comments that you can provide.

Publié : 19/06/2018 8:22 am
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

Also - Forgot to post a couple of videos of a 2x2x2 Makers Muse lattice torture test that failed and the gear bearing that I referenced in my post...

2x2x2 Lattice Cube:

Gear Bearing:

Publié : 19/06/2018 8:29 am
ntdesign
(@ntdesign)
Reputable Member
Re: i3 mk2/s MM 1.0 - random layers not extruding

Back when I still had my MMU I've seen this with feed path blockages. That's why all colors are affected, it can't push any filament through for a while. Then suddenly the pressure gets high enough to release the blockage and it just continues.

It can be highly reapeatable, I sometimes got the issue on exactly the same layer(s) on repeat prints. Possibly this kind of blockage results from some issue with the feeder gears, not the hotend itself (this kind would not usually "fix" itself and you would find blobs or overextrusion somewhere). Maybe a tiny bit of filament isn't properly retracted due to slippage and blocks up the path. I would start with cleaning filament from the gear teeth and checking spring tension.

Of course it's possible your board, wiring or relay(s) cause problems, but afaik that hasn't been reported before in this context.

EDIT: I just remembered I had all sorts of weird issues whenever I flipped on quiet mode. Just make sure it's set to off or auto before you proceed.

Publié : 19/06/2018 9:45 am
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

Thanks for replying!

I've checked the feed gears already, and all appear to be clean. What's frustrating is that my issue is not easily repeatable.

Also, I've recently run through the cold pull process, and had almost nothing in the hot end, I don't *think* it's a blockage or dirty, but the 4-to-1 combiner thing above the hot-end does have a tendency to be problematic and jam for many folks. I'll take it all apart again in case I missed something and report back.

Also, quiet mode is off. Good catch though! That was awful to troubleshoot (because I forgot I had turned it on...)

BTW, the first test print with the extruder 1 directly plugged into RAMBo-mini was successful, I just started another print that I'll check in the morning.

Publié : 19/06/2018 9:52 am
ntdesign
(@ntdesign)
Reputable Member
Re: i3 mk2/s MM 1.0 - random layers not extruding

OK, looking at your post it seems the extruders are switched by somehow combining relay states (only 3 relays, but 4 motors).
Maybe one of the relays is failing. Seems a bit early but these things can happen.
Maybe you can narrow it down by finding one or two extruders which don't fail (the one(s) which are triggered with the failing relay in closed state if that makes sense). Or just get three new ones and replace all, I don't see any other components on your images I'd expect to fail (did you notice any PTC fuses? My board is buried in a box in the basement...).
It's kind of interesting actually, I never thought about it but now wonder why they use 3 relays. Shouldn't two be sufficient? E.g. 0,0->ex1 / 0,1->ex2 / 1,0->ex3 / 1,1->ex4. Power would anyway just be passed through from the Rambo.
I'd keep in mind that it could still be "regular" step skipping due to mechanical issues somewhere, dodgy power supply and all those things. But if you're good with the Rambo directly connected you can exclude that.
Checking for blockage in the MUX is certainly a good idea, especially check the metal tube edges for chaff and all Festos for movement (which can allow the filament to relax too much).
One more thing: Filament diameters can be unevenly distributed across the spool, even with original PR filament. If you get a slightly thicker section, friction will increase in the Bowden and metal tubes and lead to underextrusion or partial blockage. I used to do a short difficult test print for each spool first (and again after using up half the spool), then just don't use the problematic ones. About 1.69-1.72 is ok. There's not really anything you can do about it, except carefully keeping track and buying different filament.

Publié : 19/06/2018 11:16 am
PJR
 PJR
(@pjr)
Antient Member Moderator
Re: i3 mk2/s MM 1.0 - random layers not extruding

So the key here is that this is not an MMU issue if it also happens with single filament prints.

I would suggest that you first reduce retraction values; you need to get single-filament prints working consistently before MMU prints.

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…

Publié : 19/06/2018 2:44 pm
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

nils.t: OK, looking at your post it seems the extruders are switched by somehow combining relay states (only 3 relays, but 4 motors).
Maybe one of the relays is failing. Seems a bit early but these things can happen.

Based on my research and findings, Extruder 1 is directly connected to the RAMBo-mini board when the MK42 Stepper switch is in a "default" state - meaning that the relays are only in use for Extruder, 2, 3, 4.

In the test print of the multi-material gear-bearing, all 4 extruders appeared to fail at the same layer height.

nils.t: I'd keep in mind that it could still be "regular" step skipping due to mechanical issues somewhere, dodgy power supply and all those things. But if you're good with the Rambo directly connected you can exclude that.

So I have a new interesting symptom which makes me wonder about your point... After I plugged Extruder 1 directly into the RAMBo-mini board and started a print - I did notice that the Bowden gears appeared to be jumping on the filament, but when I looked very closley I realized that the bowden gears were NOT skipping, but rather the stepper motor was actually missing steps. This was quite surprising to me - I've only ever seen this when the hot end jammed - yet my recent test prints appear to be extruding correctly.

You make a very interesting point about the filament quality halfway through the role. Another thing I didn't think about.

Peter: I would suggest that you first reduce retraction values; you need to get single-filament prints working consistently before MMU prints.

I might be a little confused here. What do you think the reduction of retraction values is going to do to help when we think it's filament failing to come out of the hot end? Maybe it's retracting too much and jamming??

Publié : 20/06/2018 4:07 am
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

UPDATE ON TEST PRINTS:
I've printed 3 gear-bearings with the Extruder 1 directly connected to the RAMBo-mini board.

So far, all have been a success... I'm starting a 4th tonight.

Publié : 20/06/2018 4:10 am
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

The 4th print failed. I could see that no extrusion was happening, but was able to tell that the stepper motor was at least attempting to move the filament because you could see it coming in and out of the back of the extruder, as if it was actually printing. When the print finished, the extruder seemed to move quickly and ended up squirting a bit of filament out of the hot end. I took a look at the filament and it appears to be a little bit ground up and has a couple of pockets of ground away material - so this might indicate that there is a mechanical issue, like everyone seems to be thinking.

I've taken the hotend out and apart and fully cleaned everything, re-assembled calibrated XYZ and am about to set the z adjustment again, then start test printing again.

Publié : 20/06/2018 8:59 am
ntdesign
(@ntdesign)
Reputable Member
Re: i3 mk2/s MM 1.0 - random layers not extruding

Is it just the picture or are the gears fused ?
What Peter means is that some (actually many) filaments don't work very well with the default retraction settings, too many retractions can ultimately lead to blockage. Frustratingly, sometimes this happens many hours into prints.
I had some step skipping during long/fast retracts between color changes, but those shouldn't happen in single prints. You can print the bypass to exclude friction in the MUX during testing: https://www.thingiverse.com/thing:2736405

Publié : 20/06/2018 9:37 am
PJR
 PJR
(@pjr)
Antient Member Moderator
Re: i3 mk2/s MM 1.0 - random layers not extruding


Peter: I would suggest that you first reduce retraction values; you need to get single-filament prints working consistently before MMU prints.

I might be a little confused here. What do you think the reduction of retraction values is going to do to help when we think it's filament failing to come out of the hot end? Maybe it's retracting too much and jamming??

Just don't over-think the problem. It is very unlikely to be a printer hardware issue. Run the printer with MMU connected, simplify the feed path if you have any fancy contraptions on the filament feed, reduce the tension on the feed pulleys to a minimum and reduce the retraction in the slicer.

Also, ensure that the Bowden tubes are held firmly in place by the Festos (no movement in/out of Festo).

Print a model which requires many retractions (Maker's Muse model?); make sure you can print it in single filament mode.

If you use KISS, you can prepare a model to minimise retraction values or if you are using Linear Advance (and have calibrated the correct value - which you can do with KISS), you should be able to set retraction values to 2mm or less in Slic3r.

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…

Publié : 20/06/2018 1:00 pm
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

Thanks Nil.s and Peter.

Asking me not to over-think the problem is the exact opposite of what I do for a living and in my hobbies! This being said, I'm pretty sure I've gone down a wrong path with the electronics issue...

I've been doing a lot of reading about random print failures, and Nil.s touched on it as well... The hot end can get clogged, and it, somehow, appears in randomly failed prints. I can not explain this but (trying not to over-think it) I went through the process of cleaning my hot end the other day and I've had 6 successful prints in a row, all using the multiplexer quad tube combiner thing (although I was getting ready to print up the single one). Ever since I cleaned the entire path and paint-gun cleaned the nozzle, I don't appear to be getting extruder skips and the models are printing successfully.

I'm still going to do a number of tests through each extruder directly connected to the RAMBO-mini, because I have the time and filament to fully troubleshoot this - but if these next set of prints are totally successful then I'll add the MK42 switch board back into the mix and go from there.

Do you guys have any suggestions for a print spool box of sorts? I wonder if dust accumulated somewhere in the hot end that caused this to suddenly happen...

Publié : 22/06/2018 8:07 am
FaultyLine
(@faultyline)
Trusted Member
Topic starter answered:
Re: i3 mk2/s MM 1.0 - random layers not extruding

UPDATE:

The printer is printing just fine now. Fortunately, it had nothing to do with the multi-material electronics or miniRambo board and had everything to do with a clogged hot-end and filthy filament. I'm a bit embarrassed that my whole problem was an amateur one - but at least I now know... 🙄

After taking apart the hot end, baking the individual pieces with a paint-stripping heat gun and reassembling the whole deal AND building a very crude filament filter - everything has been working fine for days now.

Here is the microfiber cloth I'm pinching on the filament before it goes through the printer:

Thanks to everyone for all of the helpful tips and not making me feel stupid for making a stupid mistake 😀

Consider this one CASE CLOSED!

Publié : 04/07/2018 10:34 pm
PJR
 PJR
(@pjr)
Antient Member Moderator
Re: i3 mk2/s MM 1.0 - random layers not extruding

🙂 My printers are in a very dusty environment; I usually change the nozzle every 4 to 6 weeks and I use a piece of sponge from a fish tank filter to wipe the filament as it enters the extruder drive system.

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…

Publié : 05/07/2018 10:03 am
Partager :