What needs tweaking? ramming?  

Re: What needs tweaking? ramming?

Sheep MMU stl from examples.

I've done the bowden tube length calibration. 1st color comes out pretty good.

But as more and more color changes occur there seems to be a buildup of filament strings or filament tips gets too big to fit back into hotend - i dont know for sure, new mmu user. Havent had a successful mmu print yet. Last failure occurred at extruder. Had to remove orange bowden tube, remove filament sensor cover and remove jammed filament from extruder. Pic below. Gonna try higher filament temp.

First question: Have you successfully printed one of the pre-sliced gcodes that came with Slic3r? There's gcodes which take from three to 35 hours, so you can give your MMU2 a solid workout using a gcode that's supposed to work.

I've been printing successfully for less than 24 hours, but one thing that's already clear to me is that there are far more variables involved with the MMU2 than single material. If you can successfully print one of the test gcodes, that eliminates a bunch of possible problems.

Re: What needs tweaking? ramming?

Ok ok, I was about to yes I printed same sheep you did. But then I read your response closer....... I did slice the model myself. I will try the presliced gcode now, thanks.

Re: What needs tweaking? ramming?

Generally, ramming parameters are determined by the tip on unload.

Honestly, I have a job understanding Slic3r settings for this. Basically, I ram as fast as possible; the ram volume determines the shape of the tip:

Less ramming - short tip plus and more stringing
More ramming - longer and thicker tip and less stringing

13 mm^3 is about right for most PLA, 15 mm^3 seems better for BVOH

Peter

question 1
Im making progress. After 4 failures, completed the provided example sheep gcode. Not my best print, but at least not a total reset printer give up and try again. Stringing is leading to failures/jamming for me.
I'm definitely having issues with stringing on supplied gcode example (see pic in post above this 1). I dont know what ramming settings the example gcode is using. Seems like I need to go with a higher ramming setting. But the ramming settings window presents as a graph - do u mean change all points in ramming graph to 13 mm^3 for PLA?

question 2
Have issues with filament grinding at the mmu2 unit during removal of filament for a color change. My gap between idler body and pulley body is enough for filament changes to occur successfully, but sometimes the filament grinds rather then pull out. Usually filament is not stuck in extruder mk3. when this happens. What else should I try to adjust?

Re: What needs tweaking? ramming?

question 1
Im making progress. After 4 failures, completed the provided example sheep gcode. Not my best print, but at least not a total reset printer give up and try again. Stringing is leading to failures/jamming for me.
I'm definitely having issues with stringing on supplied gcode example (see pic in post above this 1). I dont know what ramming settings the example gcode is using. Seems like I need to go with a higher ramming setting. But the ramming settings window presents as a graph - do u mean change all points in ramming graph to 13 mm^3 for PLA?

question 2
Have issues with filament grinding at the mmu2 unit during removal of filament for a color change. My gap between idler body and pulley body is enough for filament changes to occur successfully, but sometimes the filament grinds rather then pull out. Usually filament is not stuck in extruder mk3. when this happens. What else should I try to adjust?

Q2 is dealt with in my "troubleshooting" thread; you most likely need to increase tension (after calibrating MMU).

Q1 is most likely the result of your Q2. Skipping on unload can cause stringing 🙁

Peter

Re: What needs tweaking? ramming?

Ok thank you...

Re: What needs tweaking? ramming?

I'll try your suggestions 1st.

Need clarification on ramming:
Less ramming - short tip plus and more stringing
More ramming - longer and thicker tip and less stringing

13 mm^3 is about right for most PLA, 15 mm^3 seems better for BVOH

The ramming settings window presents as a graph - do u mean change all points in ramming graph to 13 mm^3 for PLA?

Re: What needs tweaking? ramming?

The ramming settings window presents as a graph - do u mean change all points in ramming graph to 13 mm^3 for PLA?

Now, this is where we start to have problems 🙁 Not being a regular user of Slic3r, it's a bit difficult to comment here. All I can tell you is that I set ramming to 13 mm^3 in the slicer I use and it does the rest. I believe that the Slic3r PLA default is 13mm^3 and should be fine for most makes of PLA. You could try setting the ramming time to 3 seconds, which generates 18 mm^3 (about the upper limit for PLA), or somewhere in between. You will get a longer and thicker tip plug.

Peter

Re: What needs tweaking? ramming?

I think i didnt understand what u were trying to tell me b4. 🙄

I'll leave the graph alone. Users dont edit points like i thought. Manipulating the "total ramming time" then increases or decreases the volumetric speed mm^3/s on the graph points.

so default is 2.50 sec = 20 total rammed volume mm^3

OK so at this point I have to slice my own model so I can start tweaking ramming values. I have no idea what ramming value the supplied sheep gcode uses.

I should do the MMU calibration for the bowden 3 successive times perfect 1st. Reminds me of calculating steps for extruder.

Re: What needs tweaking? ramming?

According the the PLA defaults I have, 2.5 seconds = 13 mm^3 of ramming:

You can move the points a little to increase this to 15 mm^3; you need the speed. I ram at a constant 12 mm^3/s, with the heater turned off. Although this is above the V6 maximum, it works quite nicely and produces good tips.

Peter

Re: What needs tweaking? ramming?

Im making real progress with the supplied sheep gcode combined with filament prusa tested. (silver prusa, black hatchbox)

Only 1 jam towards end of print 1st time using Prusa filament.
2nd try, no jams entire print. (So relieved, frustration level nominal)

Cant figure out why I'm getting layer skips in the body color. I thought it was related to the jams, but now that I've printed an entire print without a jam I'm not sure why its skipping at the same point every print. Maybe my gcode got corrupted, because peter.l22 was able to print it perfectly. I dont know just happy no jams.

Re: What needs tweaking? ramming?

Cant figure out why I'm getting layer skips in the body color. I thought it was related to the jams, but now that I've printed an entire print without a jam I'm not sure why its skipping at the same point every print. Maybe my gcode got corrupted, because peter.l22 was able to print it perfectly. I dont know just happy no jams.

Glad to hear you're making progress.

The skipped layers are because the filament didn't fully load and the printer didn't detect it. I've had this happen a number of times, and it often seems to clear on its own but it leaves visible defects.

What seems to be going on is that the MMU loads the filament, but doesn't quite get it far enough down to engage the Bondtech gears. Then it unloads normally when it's time to change materials again.

My theory is that you sometimes get a slightly misshapen tip such that the point where the FINDA detects the end of the filament is too narrow for the Bondtech gears to grab the filament when it arrives next. After a few times back and forth through the system, the tip can get mushed up enough that it finally grabs and the problem clears.

One idea I have for fixing this (I don't know if this is possible given the MMU's electronics) is to have the MMU and extruder drives run synchronously for 1cm or so at the point of the load cycle where the filament enters the Bondtech gears. The current design has the MMU stop and disengage before the extruder starts, but running the two synchronously for a short distance would allow the MMU to push the filament a little further in case the extruder doesn't engage at the point where it's expected to.

Re: What needs tweaking? ramming?

It's usually caused by a string.

Either at the end of the tip, folded back and getting stuck in the orange tube, at the lower festo or top of the extruder, or by a string from the previous load (these can be up to 50cm long 🙁 )

If you are really unlucky, the filament can be ground down by the MMU drive gear and then not load for a few layers, or you could get a few consecutive failed loads.

One solution is to increase the Bondtech tension a little. Strings are usually caused by the initial retraction not operating correctly. There's a 15mm retraction at F6000 with acceleration set to E8000. And this pulls a vacuum in the hot end. If the tension is not sufficient, you will get filament skipping.

Peter

Re: What needs tweaking? ramming?

It's usually caused by a string.

Either at the end of the tip, folded back and getting stuck in the orange tube, at the lower festo or top of the extruder, or by a string from the previous load (these can be up to 50cm long 🙁 )

If you are really unlucky, the filament can be ground down by the MMU drive gear and then not load for a few layers, or you could get a few consecutive failed loads.

One solution is to increase the Bondtech tension a little. Strings are usually caused by the initial retraction not operating correctly. There's a 15mm retraction at F6000 with acceleration set to E8000. And this pulls a vacuum in the hot end. If the tension is not sufficient, you will get filament skipping.

Peter

Increase tension which bondtech gear, mmu or at extruder?

Such a tricky tension to get right. MMU2 idler body can't be too tigher or it won't rotate inside mmu. To loose and it won't push filament, to tight it will grind filament on retracts.
Same with extruder. If tight it may not release the filament during a retract. To loose and it may not grab filament during a feed. And once the filament passes the sensor, it thinks filament is loaded correctly. I wonder if filament sensor in extruder is checked during retractions....

Re: What needs tweaking? ramming?

Increase tension which bondtech gear, mmu or at extruder?

There's only one Bondtech gear; the others are just simple drive gears with idler bearings.

https://www.bondtech.se/en/product/drivegear-kits/

Peter

Re: What needs tweaking? ramming?

If at 1st you dont succeed, try, try , try, try again.

Black Sheep Success......

Supplied sheet gcode
Stinging @ bondtech extruder still caused jam, which lead to layer skip even on firmware 3.4.0 final.
Prusa - silver
Hatchbox - black

Modified slice settings --> my 1st success, no jams, no layer skips
1st layer temp 215
others 200
ramming total time 3.50 = 24mm^3 total rammed volume
Prusa - silver
Hatchbox - black

Gonna try the 3 color lizard tomorrow

Re: What needs tweaking? ramming?

Black = Hatchbox recommend printing temp 180c - 210c
Silver = Prusa recommend printing temp 200c - 220c
Red = eSun (fire engine red) recommend printing temp 205c - 225c

Slicing settings that I believe helped to eliminating stringing:
1st layer temp: 215c
other layers: 195c

Number of cooling moves: 3
Total ramming time(s): 3.50
Total rammed volume (mm^3): 24

Re: What needs tweaking? ramming?

Please upload your issues to this ticket so prusa can work on them

https://github.com/prusa3d/Prusa-Firmware/issues/1163

By the way
Did you have had luck with the prints and do they work without any issue now?

Re: What needs tweaking? ramming?

Please upload your issues to this ticket so prusa can work on them

https://github.com/prusa3d/Prusa-Firmware/issues/1163

By the way
Did you have had luck with the prints and do they work without any issue now?

I'd say im getting better at troubleshooting. The longer the print the higher a chance of problems. I'll take a look at your github issue, and add my two cents.
Trouble free printing score / slice it and forget it:
MK3: A+
MK3 & MMU2: D-

Re: What needs tweaking? ramming?

Trouble free printing score / slice it and forget it:
MK3: A+
MK3 & MMU2: D-

😆 I think it's going to be a long time before we get to "slice it and forget it" with the MMU (or any other multimaterial system). There's just too much that can go wrong, and we're still too early on the experience curve.

But I think we could get to "slice it, tweak a few things, watch the first layer carefully and then check on it every few hours." On a good day my MMU is about at that level.