First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.  

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

So you got it printing which means the sensors are most likely not causing problems.

Is it extruding normally when you "load to nozzle" the filament or can you hear the gears in the extruder grinding on the filament or clicking when you select "no" and the extruder continues to extrude? If so, you will have severe underextusion and therefore weak adhesion of every layer including the first...

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@robin

I think the problem is where the clear tubes out of the buffer, connect to the MMU. If I remove those tubes, and place a straight piece of filament in, it goes through smoothly. Once I put in the tubes and push filament through, it binds up a bit, and I think that is causing a pinch that makes it harder for the extruder to pull filament, causing grinding and such.

I'm not sure how to fix that, and align the tubes correctly. I think its here in step 13-14.
https://www.help.prusa3d.com/en/guide/6-electronics-and-mmu2s-unit-assembly_44728#45264

The PTFE tubes do not bind filament when passing through alone, and nothing binds when pushing filament without the PTFE tubes installed, but installing the PTFE tubes binds on 2 and 5 specifically no matter what I do.  🙁

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

I keep hearing of this issue, and wondered if it could be friction that was causing it, but I've never seen it myself.

My MMU didn't work properly to start with, and I gave up on it for a number of months. Recently, I decided to sort the problems or throw the thing away.

The first thing I did was print off the official upgraded part, then a replacement for the part gripping the PTFE tubes as they enter the back of the MMU body. Using Festo-style couplers helped a lot in reducing friction (the original set-up gripped so tightly that the tubes were deformed). I also slightly filed out the holes where the tubes entered the body-proper as they were a really tight fit; the couplers stop those holes needing to be so tight.

Try this;

https://www.thingiverse.com/thing:3233579

I still wasn't happy with the way the PTFE tubes were gripped at the buffer end, so I made some of these (the new one for the buffer-MMU side, the original for the buffer-spool side - I didn't end up using this one as I figured a bit of friction there would help stop anything being fed back out) they also help a lot with feeding in filament with your fingers in the first place (note: I printed two for the buffer-MMU side as this keeps the original geometry of how the filament feeds through the buffer);

https://www.thingiverse.com/thing:3545935

Finally, and just for tidiness;-)

https://www.thingiverse.com/thing:3166394

My MMU has worked perfectly for over a month now, and I've not intervened a single time (apart from PVA:-(

All the best.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

Posted by: @stewartr

I keep hearing of this issue, and wondered if it could be friction that was causing it, but I've never seen it myself.

My MMU didn't work properly to start with, and I gave up on it for a number of months. Recently, I decided to sort the problems or throw the thing away.

The first thing I did was print off the official upgraded part, then a replacement for the part gripping the PTFE tubes as they enter the back of the MMU body. Using Festo-style couplers helped a lot in reducing friction (the original set-up gripped so tightly that the tubes were deformed). I also slightly filed out the holes where the tubes entered the body-proper as they were a really tight fit; the couplers stop those holes needing to be so tight.

Try this;

https://www.thingiverse.com/thing:3233579

I still wasn't happy with the way the PTFE tubes were gripped at the buffer end, so I made some of these (the new one for the buffer-MMU side, the original for the buffer-spool side - I didn't end up using this one as I figured a bit of friction there would help stop anything being fed back out) they also help a lot with feeding in filament with your fingers in the first place (note: I printed two for the buffer-MMU side as this keeps the original geometry of how the filament feeds through the buffer);

https://www.thingiverse.com/thing:3545935

Finally, and just for tidiness;-)

https://www.thingiverse.com/thing:3166394

My MMU has worked perfectly for over a month now, and I've not intervened a single time (apart from PVA:-(

All the best.

Thanks for this, I will get moving on this stuff now.

It's a CONSTANT GAMBLE whether or not the MMU will load or not load the damn filament. Even if the EXACT same filament port is used. I'm at my wits end, because I could not possibly expect the MMU to do a multimaterial print when it barely works as a single printer. If I had known it was going to be this much trouble, I don't think I would have done it, because it's a pain in the ass to even switch out filaments with it.

I can't believe how awesome an experience the base printer is, and what a hellish experience the MMU has been.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@lucas

I share your pain. Or rather I did:-) Those things sorted out all my issues (maybe not the tube organiser, but it keeps things neat). But I never had the problem you have, so...

You'll need 10 Festo-type couplers (~£10 on Amazon, the cheap ones are good enough here). The Prusa upgrade I printed was the selector head (modified blade holder, and slightly redesigned body).

And remember the standard mantra - *check the positioning of the optical sensor* on the print-head and make sure the MMU *spring-loaded bolts are tensioned correctly* (flush with the body, then two turns). These don't sound like the problem you're having though. (It might also help to try with Prusament filament and Prusa Slicer settings, just to eliminate another variable.)

Hope it works out - it's worth it.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

Stewartr is right. I have my MMU working nicely for about 2 month with all the upgrades stewartr mentioned and 2,5 and 3 mm ID tubes. I used PRUSAMENT PETG, Verbatim PLA and Filamentum PLA and everything worked. 
Yesterday I decided to use some older PLA (Alicia) for a single mode print, watched the first layer print perfectly and went to make dinner. After an hour or so I returned to find a failed print due to intermittent underextrusion. Tried it again after cutting the tip with the same result... so apparently the diameter tolerance of that filament is to high. Conclusion: if you have issues with the MMU, use fresh filament that is proven to work with the MMU to rule out bad filament as a cause...

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

I am using the silver filament that came with the printer for all my tests.

So I printed the first part that goes on the MMU and now I am printing the buffer parts as a bench test for each port.

First print is going well right now, but the filament didn't load to the nozzle the first try, it still retracted it once and tried again.

Is there anything at the extruder level I should be looking at?

For instance, maybe I have the tightness of the extruder gears too high or low? What else could be going on?

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@lucas

To print the replacement parts / 'improved' parts without MMU issues, you already seem to have the solution - remove the PTFE tubes. Although that'd be a pain because then you'd have filament feeding problems unless you could position the spool directly behind the unit. When I printed these I'd already had my MMU disconnected (@jsw has a nice mod' to make this easier, but I just removed the power and data cables (carefully) and unhitched it).

As for the extruder gears - that's just the normal MK3 fiddling - it they, or the motor, make clicking sounds then adjust the tension. Can't help much there I'm afraid; I do it 'by-ear' / 'gut-feel' - too tight does make a slightly different noise from too slack, but it's a subtle difference. Normal advice is the bolt end flush with the nut, then adjust for softness of the filament. Like I said - not really much help.

BUT - *remember the standard mantra*! Those two things are the cause of ~90% of issues (assuming the thing worked at some point).

You sound like you're on top of things though. I'm surprised fitting Festo-couplers on the MMU body didn't help that much - that was my main issue. The buffer-side mods just made things a bit better and easier. It depends where you're getting all the friction problems though.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@robin

I've read about people using larger internal diameter PTFE tubes back-side of the MMU; does that help much?

All the filament I use isn't going to be varying by *that* much, and I learned the hard way to not try and pull used filament out the back before ejecting it (out the front) first and cutting the end - this also saves me the hassle of having to cut it when it's back on the spool.

I'm intrigued...

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@lucas

I've just remembered the grief I had with the buffer-end PTFE tubes, so maybe that mod' will actually help more?

I'd put the tubes in and tighten the bolts (a really crude mechanism) to grip them (creak, creak, squeal). A few days later one of the tubes would get pushed out so I'd loosen that bolt (creak, creak, squeal), put the tube back in and tighten it further (creak, creak, squeal, squeal). Repeat a few days later...

That must have introduced significant squeezing! Still - my main problem was the MMU side.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

Another day another problem, and I'm beginning to think that Friction isn't it.

I did a bench test printing on each port of the modified tube holder, and they all seemed to go pretty good.

However, I tried to do a print with PETG and the printer responds in a totally different way, and I don't understand why. When I "Load to Nozzle", it pushes the filament out in short bursts and keeps pushing even when its grinding.

This filament I used for about 200 hours printing the parts for the Ikea Lack Enclosure, so I don't think its a filament problem.

1. Why would it treat that a different way

2. Is that behaviour a symptom of something else?

I am actually starting to think that I have issues with the filament sensor. Anyone have resources for troubleshooting this?

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

I rarely load to nozzle nowadays, so I can'r remember what steps it takes in detail. But the general operation is that the MMU pushes filament for a set distance (the distance between the FINDA probe in the MMU and the filament sensor in the extruder) and then disengages the MMU drive. All the while the extruder gear is spinning ready to grip the filament. If this, for some reason, fails then the MMU retracts back to the PINDA position and tries the same manoeuvre again. At some point it stops (perhaps if it never gets a signal from the controller that filament has been detected at the extruder?) then and feeds in short bursts (about an inch) with a little pause between them. I think the idea is to see if the extruder will take the filament. I'm sure someone else can explain this more accurately and correctly;-) Part of this sounds like what you describe. (I'm describing from memory what I hardly ever see, so forgive my vagueness.)

There are two things I can see going wrong here; a) the extruder gear tensioning bolt is mal-adjusted to it doesn't grip the filament, or b) the filament sensor on the extruder has worked loose and doesn't detect the filament. *Remember the standard mantra*. Go to the menu 'Support/Sensor info' and look at the IR sensor value. With the PTFE tube between the MMU and printer disconnected at both ends push a bit of filament in and the value should change (takes a bit of force and you have to push the gear apart to trigger the sensor). You could try pushing the sensor housing back and forth to see if this causes any change. Just to be sure, do similar with the FINDA sensor. If it's fine with PLA, but doesn't work with PETG then I'd be inclined to suggest the tensioning bolt. But try all checks.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

Posted by: @stewartr

I rarely load to nozzle nowadays, so I can'r remember what steps it takes in detail. But the general operation is that the MMU pushes filament for a set distance (the distance between the FINDA probe in the MMU and the filament sensor in the extruder) and then disengages the MMU drive. All the while the extruder gear is spinning ready to grip the filament. If this, for some reason, fails then the MMU retracts back to the PINDA position and tries the same manoeuvre again. At some point it stops (perhaps if it never gets a signal from the controller that filament has been detected at the extruder?) then and feeds in short bursts (about an inch) with a little pause between them. I think the idea is to see if the extruder will take the filament. I'm sure someone else can explain this more accurately and correctly;-) Part of this sounds like what you describe. (I'm describing from memory what I hardly ever see, so forgive my vagueness.)

There are two things I can see going wrong here; a) the extruder gear tensioning bolt is mal-adjusted to it doesn't grip the filament, or b) the filament sensor on the extruder has worked loose and doesn't detect the filament. *Remember the standard mantra*. Go to the menu 'Support/Sensor info' and look at the IR sensor value. With the PTFE tube between the MMU and printer disconnected at both ends push a bit of filament in and the value should change (takes a bit of force and you have to push the gear apart to trigger the sensor). You could try pushing the sensor housing back and forth to see if this causes any change. Just to be sure, do similar with the FINDA sensor. If it's fine with PLA, but doesn't work with PETG then I'd be inclined to suggest the tensioning bolt. But try all checks.

Thanks for this info. How do you load your filament to the nozzle then, how does it know which one to use if nothing is loaded? Is there a good place to find procedures for printing with the MMU (I don't find the manual super friendly)?

I've found that if I don't load to nozzle first, it gives me flak or freaks out. That could be Slicer issues I guess, since you have to setup which nozzle in there?

I also have set the tension screws to the recommended values I believe (Flush and 2 turns on the MMU, Flush on the Extruder).

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@stewartr

Well there was definitely something wrong with the IR sensor. I put in the filament and couldn't get it to stabilize the reading when moving around the filament/sensor cable, so I re-set the sensor in its holder and it seems to be solid. I am now printing that PETG part with no issues so far.

Fingers crossed. lol

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@lucas

That's about right, but with spring loaded bolts it can never be an exact science - your springs will be different to mine. The guidance is 'ish' / 'about'...

The MMU end (assuming minimal friction) only has to push back and forth through PTFE tubes so it's not as critical (hence flush and two turns down works well - something I read on this forum a while ago). The extruder gear though has to feed filament at vastly varying speeds and direction into, in effect, a pressure chamber. This really does need to be a good tension. And different filament (i.e. type / manufacturer) has different density / softness so needs different amounts of tension. I've kina hit a 'sweet spot' for both PLA and PETG from Prusa, but still get the occasional bit of clicking during a print, so I'm quite good at a quarter turn with the print head moving at speed... When I change the manufacturer, well - I've got a note of offsets from the Prusa setting and only have minimal tweaking to do on my first print.

The MMU / extruder gear interaction during loading is a bit different though, so getting that right is more a case of observe / adjust / observe. It takes practice, and then is almost impossible to describe. (As in 'gut-feel'.) All assuming that the extruder filament sensor is adjusted correctly, obviously...

(I tend to load up spools for a while and slice everything for that set-up. If I finish what I'm doing and change colour, or don't use the printer for a week or so and put the spools away, then I have to re-slice. The chances are that either the firmware or the slicer will have been upgraded before I print the same thing again, which needs re-slicing anyway. I figure it's a minimal inconvenience, and works for me.)

Which brings me to an obvious point; you are using the latest firmware / slicer???

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@lucas

Good news! Hopefully all sorted. That will work loose over time, and it has to be tightened 'enough so that it holds in place but not enough that it breaks'.

Don't you just love instructions like that!?!?!?!

BTW: @joantabb has a good hack for the filament sensor, if you're good with a soldering iron.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

@stewartr

I am one step behind on the firmware (3.9.3) which I will update today.

I'm pretty good with a soldering iron, and have an extra sensor (long story) if you wanna link that mod.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

I'm afraid I can't remember; it was one of the MMU forums though and it crops up with every issue that's down to the filament sensor not being adjusted correctly. She soldered a LED across two pins and drilled a hole in the top of the cover. You'll find the instructions at a glance because there was a photo of a LED.

I really should upgrade to 3.9.3 myself.

RE: First Layer Calibration: MMU2S constantly loads and unloads the filament to the nozzle.

You can just add an indicator LED to both of the critical sensors. Either just put it between singnal and ground (with respective resistor) or split it with a transistor if you don‘t trust the constant current... I have a green light on if FINDA detects filament and a red light off when filament reaches the extruder IR Sensor. Both with split signals and both together next to the MMU so it‘s like a little traffic light.

The IR signal is 0 when triggered, the FINDA is 1.

Here are two possible mods: 

https://www.prusaprinters.org/de/prints/2139

https://www.prusaprinters.org/de/prints/29619-prusa-mmu2s-finda-troubleshooting-aid