Globs and Slight Spattering from Nozzle in Wipe Tower
Just got my MMU2S added to my i3 MK3S and so far one step at a time the process has moved along. I am however currently stuck on a problem I cant figure out. I have however noticed a couple things that might all be related.
The main issue is globs of filament being dumped into the wipe tower after each filament change as can be seen in the attached image.
Now, what else Im seeing that might be causing this behavior.
- As the filament enters the extruder from the MMU2S, there is a quick and slightly chunky sound in the bondtech drivegears. Almost like they are pulling faster than the MMU2S is pushing the filament down the PTFE tube causing them to grind the tip of the filament just before it actually grabs.
- Once the filament passes the bondtech drivegears and is fed through the heatbreak and nozzle it almost seems like its maybe going into the nozzle too fast as it takes a second longer than I feel it should to come out and it all a sudden shoots out a bit as though it was building up a pressure behind it. This is what is causing the globs in the wipe tower.
First, Im thinking it might be two issues, but the globs in the wipe tower Im wondering if that is an issue with the extruder not allowing enough time for the filament to heat back up inside the nozzle before pushing it through. Is there a setting to slow that feed rate down?
Second, the slight grinding noise as the next loaded filament enters the bondtech drivegears, maybe that could be fixed by either slowing down the bondtech drivegears or speeding up the MMU2S feed into the PTFE tube?
Or, perhaps these two issues are directly related to the bondtech drivegears just moving too fast?
Thoughts, ideas? Anyone seen this before?
Just wanted to add a few things I have tried:
- I have checked the filament tips coming out of the MMU2S after a change and they do not appear to have a bulbish head nor are they very stringy. I thought maybe that might account for the bondtech drivegears clicking.
- I have tried reducing the print speed down to 50% to see if that would allow the filament more time in the nozzle to heat up in case that was the problem. Seems to have not made any difference as the bondtech drivegear clicking and the pressure in the nozzle/filament shooting out causing a glob in the wipe tower still happened.
- I have tried increasing the nozzle temperature up to 20 degrees more than I normally print PLA and that might have made a little difference as I think the "popping" noise of the filament bursting out of the nozzle seemed to have reduced a bit. But raising the nozzle temp increases the risk of stringing on the filament tip after unloading as well as potential heat creep.
I'm a Mini owner and a long time MK2 owner and I just finished building the MMU2S and upgrading my MK3 to a MK3s in the process. Besides resolving a slight interference issue between new extruder and the connector housing on the back of the bed (did you check this BTW?) it went flawlessly and works magnificently. I saw your post and wanted to comment on your observations:
"Im wondering if that is an issue with the extruder not allowing enough time for the filament to heat back up inside the nozzle before pushing it through."
No - not if everything is working properly.
"As the filament enters the extruder from the MMU2S, there is a quick and slightly chunky sound in the bondtech drivegears. Almost like they are pulling faster than the MMU2S is pushing the filament down the PTFE tube causing them to grind the tip of the filament just before it actually grabs."
Mine doesn't make any sound. As the extruder's Bondtech gears catch it the spring-loaded panel bumps out slightly (which drives the fin into the IR sensor and should signal filament to the Einsy). Remember the time overlap for push/pull between the MMU2S and Extruder is very short; as soon as the sensor in the extruder signals that there is filament present then the MMU2S disengages. That said, have you re-checked that the extruder IR sensor is calibrated correctly? This is very sensitive and many problems stem from this being adjusted wrong. Calibration is described in the start-up instructions and involves using the small Allen key being inserted into the Bondtech gears.
"Once the filament passes the bondtech drivegears and is fed through the heatbreak and nozzle it almost seems like its maybe going into the nozzle too fast as it takes a second longer than I feel it should to come out and it all a sudden shoots out a bit as though it was building up a pressure behind it. "
When mine does a filament change the extruder gear movement is pretty interesting, I can't claim to know exactly what it is doing but when it's loading it does seem to do what you describe above without any issue. On mine the MMU2S pushes it to the extruder, the extruder "catches" it, then it seems to feed very quickly with nothing coming out. (It seems to me at this point perhaps the nozzle is largely empty and it's filling the nozzle cavity but IDK). Then it (strangely) retracts, then it feeds again. If it's at the very beginning of the print (i.e. starting the purge strip on the bed) then the first thing I get is a small round "blob" and then the extruder moves to make the normal purge line. If it's during a layer change then after loading the new filament the extruder starts making slow passes over the wipe tower with nothing coming out; it takes about 5 seconds before the old color filament start coming out, 5 more seconds and then the new color.
So that last part is important - after a filament change the extruder starts moving like it's laying down a line over the wipe tower with nothing coming out, and slowly filament comes out - no blob.
"I have tried increasing the nozzle temperature up to 20 degrees more than I normally print PLA and that might have made a little difference as I think the "popping" noise of the filament bursting out of the nozzle seemed to have reduced a bit. But raising the nozzle temp increases the risk of stringing on the filament tip after unloading as well as potential heat creep."
Obviously it should all work fine at the proper filament temperature. BTW are you using "good" filament? When I see people write about "popping" sounds it's sometimes related to filament that has not been stored in a dry environment and has absorbed moisture. The absorbed moisture changes the filament characteristics and can also make a "popping" sound from the moisture in the filament expanding.
Some things I would consider:
- Make sure you're using "good", dry filament.
- Make sure you have the latest firmware loaded for the MMU2S, the MK3s, and are using the latest Prusa slicer.
- Check that the IR sensor is calibrated correctly.
- Are you sure the extruder Bondtech gears were properly aligned with the filament (such that it grabbed the filament correctly)?
- Swap the nozzle (in case there is blockage in there).
I know it's frustrating but don't give up - when you get it working you'll love it!
Thank you Steve for taking the time to read through my post and offer up some suggestions! Let me add a few notes in response:
- I forgot to mention in the OP, this MMU2S was added to an i3 MK3S that has been running perfectly for the last year, so I at least know I can rely on the fact that I started with a good working printer. Also I have never had popping filament before and have printed a couple dozen different brands/colors/etc of PLA. I store all my filament in dry containers and regularly change out the silica gel. Also, if this were an issue with moisture in the filament there would be popping at other/random times and I am ONLY hearing that sound after a filament change and only once each time.
- I have uploaded a video showing the sounds I am hearing. They aren't picked up too well by the microphone but if you turn up the volume or wear headphones you can definitely hear them. LINK. The bondtech drivegear clicking noise can be heard at 40 seconds and the filament pushing out of the nozzle at 45 seconds.
- Regarding the IR sensor calibration. Well I could certainly go through that process again and probably will since you've suggested it, but basically I just followed the Prusa guide here (LINK) and thought it appeared to be working fine to me.
- The way you described what happens when your machine does a filament change makes me at least feel like the behavior of mine is at least mostly normal. Mine does do the same on the last part, where it will move like it's wiping with nothing coming out and it's right there that the popping/spattering happens when the filament does burst out.
- Regarding filament, well I'm really not sure what "good" filament is these days as I've tried about 8 different brand and they have all performed well for me. The brands in particular I have been using with my MMU2S setup are Hatchbox and 3D Solutech. Both of which I have used a lot of and never had any issues.
- Latest firmwares (3.9.0 and 1.0.6) are loaded onto the i3 MK3S and MMU2S.
- I have not checked the nozzle for a potential issue there. Didn't think there might be since the machine had been working fine. However who knows whats gone down the pipe since adding the MMU2S. You recommend trying a brand new nozzle or a cold pull?
I really wish Prusa would makes some simple videos just showing a detailed example of what normal printing should look like with their machines. If you had never seen one of these things operate how would you know if yours it working properly?!
And don't get me wrong, I am not going to give up on the MMU2S, I only just got started. And it IS working as I've already gotten a couple very simple test objects printed and had not issues other that those listed in this post. I'm just trying to understand whats going on here as I like my machines to run as good as possible and to understand as much as I can as to how they work.
You know what Im wondering now... What if the extruder is having trouble pulling the filament from the spool through my buffer system. I am not using the stock Prusa buffer. I found it to be way too bulky. So I built this setup: LINK. It works perfectly but for all I know it might create too much resistance along the filament path causing the extruder motor to struggle. That would explain the filament not coming out evenly all the time and the clicking noise when the bondtech drivegears skip. Perhaps I'd need to upgrade to a geared ratio extruder? Like the Bondtech 3:1 (LINK)? But would that extruder even work with the MMU2S chimney IR sensor setup?
Nevermind, it looks like they made a new part to deal with the chimney IR setup that the MMU adds:
To use the MMU2s together with the Bondtech an extra Add-On is required. This will allow sensor to trigger once the filament is reaching the drive gears.
You can download the STL file bundles from our MMU2S Add-On Files tab.
All credits to Leon Van Roosmalen for his amazing design. Find more information about his work and designs here: https://www.thingiverse.com/Leonrsmln31/designs
So the extruder gear motion looks perfectly normal, and the sounds (besides the popping) all seems normal too. If the noise is coming from the extruder then yeah it kinda sounds like the Bondtech is skipping or having trouble with the filament. I didn't realize you had a custom buffer and feed system, which could be a problem. A simple test is to feed 2 filaments on the stock rollers directly into the MMU2S PFTE tubes, no buffer at all. Just manually rewind the spools while you're watching.
Steve, that was a great suggestion and I should have thought of that one. So I got three spools on holders directly behind the printer and loaded the filaments right into the MMU2S, withouth even using PTFE tubes into the back. So that was the least amount of friction this system could have and it made no difference. Exact same globs in the wipe tower and the extruder bondtech drivegears clicking once loaded into the nozzle. The extruder motor did seem to run a little cooler so that might be due to the reduced load from the lower resistance, but still no change in the wipe tower issue.
I think the issue is the Max Volumetric Speed and that the setting for that is ignored for the wipe tower. If that speed is defaulting too fast that would account for all my issues I think. However it looks like you cant change it for the wipe tower. I have created a new thread on this issue here (LINK) and will probably submit it as a feature request in the PrusaSlicer GitHub.
Hmmm.... Okay, maybe my eyes are going cross-eyed from staring at all these settings, but after restarting PrusaSlicer it appears that the Max Volumetric Speed is getting applied to the wipe tower now. Im not sure what I did different as the setting was applied to both sections in the Print Settings and Filament Settings.
So Im next going to play with reducing that setting and see if it makes any difference.
Nope, nevermind, it's still happening. See the attached pic. That is with the Print Settings and Filament Settings sections set to 5 mm3/s. Those darker orange spots are exactly where I'm getting spattering/blobs in my wipe/purge towers.
So FYI your wipe tower looks identical to mine, and the volume values (I was using PLA) are identical in my slicer. (I don't think it matters but I should note that my prints use the Wipe to Object Infill and all my observations are based on that).
Something is going on during the filament change, and the extruder clicking sounds are a clue.
I have a simple question - have you tried slicing your model with all the default values of the Prusa slicer? I know many folks that have spent many, many hours debugging an issue because they tweaked a setting in the slicer even though they are running stock Prusa hardware. As a baseline if you have a stock hotend, nozzle, extruder gears, etc. you should get perfect results with the default settings. If you are not sure about the setting (I often see folks forgetting about a change they made a long time ago that is now in their profile) I usually (strongly) advise people to simply take 10 minutes and remove, then re-install the Prusa slicer. In my years of printing on my 3 PRUSAs I have never changed any settings except for filament and bed temperatures...I'm not a fan of changing any other values that Prusa has proven to work. Almost always when people change thee other values they are working around a different issue (e.g. bad PEI surface).
I don't think it's worth going down the Volumetric settings but since you're there already, there are 2 Max Volumetric Speed settings, one in Print (an overall Max limit that never gets exceeded) and one in Filament (to adjust for different filaments). Again, these settings normally need no adjustment at all and I don't think this is your issue. But anyways which one are you adjusting?
Going back to the extruder clicking noises- have you tried adjusting the extruder's Bondtech gear tension? (Nominally the head of the single tension adjustment screw should be flush with the side of the extruder so that the head is just recessed - see assembly manual). Mine is flush with the extruder. When you're done with the print, or when it's stopped early, do the gear marks on the filament near the end look normal? Working your way up the filament about 9 inches (where the MMU would have grabbed it) are there any unusual gear marks? (If the MMU and extruder are not working together the MMU gears can eat away at the filament and cause feed problems when that spot hits the extruder gears).
When the filament comes out during a change can you tell if the tip has the "conical" shape to it? I have the clear PTFE and with a darker filament I can clearly see a well formed cone on the end. And when I extract the filament at the end of the print I see it also. This will tell you if the overall retraction sequence is properly working.
And - just for fun, have you tried feeding the 2 filaments through different ports on the MMU? I assume you're using #1 and #2...in case the MMU is somehow involved try using say #5 and #6. As you watch are the MMU rollers engaging properly? Are the MMU gears properly aligned to the filament?
Also - just for fun, have you tried a single filament "normal" print recently just to make sure everything is still working correctly?
During the filament change at the end of F1 there is a unfeed/feed/unfeed routine that I believe is shaping the end of the filament into a cone, then F1 is unloaded. Then F2 is fed in, unfed slightly, then fed in again quickly (which is where the max volume in red appears on your diagram) and then the feed rate slows down. This is where I'm a little unsure (I'd have to go back and watch the sequence carefully) but I believe on my printer during that max volume period the extruder is sweeping over the wipe tower and nothing is coming out for about 5 seconds...I think it's filling the nozzle cavity during this time. By the time the filament starts actually extruding the feed rate has been lowered back to nominal or just slightly above it and it starts extruding and filling the wipe tower again.
During that last part it looks like you are ejecting a large amount of filament at a time that I believe the nozzle cavity should be empty and filling and nothing should be coming out, OR it is empty and filling but for some reason it's blocked and pressure builds and then it comes out in one big spurt. (That's why I previously asked about the nozzle condition). As I mentioned before it's really important for the IR sensor to be working correctly here for many reasons. You can ensure it's working via 2 methods: 1) as soon as the filament enters the extruder Bondtech gears you should see the side of the extruder push out and the MMU light should turn green and the MMU should immediately disengage, conversely during extraction the moment the filament leave the extruder gears and the extruder side retracts back in and the MMU light will immediately turn red. Also 2) you can watch the sensor setting with the LCD display.
Sorry I know my response is all over the place. Also LMK if you want to end this post and switch over to your other post. Keep us updated.
Steve that was a lot to chew on! Thanks man! Had to take a couple days off from this setup to work on other stuff, BUT there was some interesting developments.
You "pure Prusa" approach of trying to get the system running using all the default settings intrigued me. For the record I did try the first couple prints at all default settings and when they didnt work thats where the modifying began. I thought about your suggestion of reinstalling PrusaSlicer, but that seemed like something I shouldnt have to do as the default settings should be easy enough to switch back to by clicking the little lock icon on all the settings that were orange/changed. However I started thinking before even doing that, what if I could try some MMU2S test .gcode prints that Prusa might have out there. So I searched, and found the ones on the bottom of this page (LINK). I first tried the 3 filament one and then the 5 and whatda know? They printed fine with no extruder clicking or globs in the wipe tower (see attached image).
So then I go back to PrusaSlicer and reset all settings back to their original values, create a simple 5 filament object (a bar .25" high made up of 5 .5" cubes), slice it, copy it to the SD card, start the print and its back to the same problem with the messy wipe tower and extruder clicking.
So a .gcode file directly from Prusa seems to work fine, but something sliced in PrusaSlicer, maybe not? Starting to think you might be right and the next step might be to re-install PrusaSlicer. Just worried all my settings and filaments might be lost.
Also to answer all the machine functioning questions you had it your last post, yes the system seems to be working mechanically as its supposed to. And that can kinda be confirmed by the Prusa test prints successfully printing. I've attached an image of the filament tips after the 3 filament Prusa test print finished. Little stringy on the white one but usually they look like the other two in all the tests Ive done except for the Hatchbox blue filament Ive got. For some reason that color is very stringy. All of the colors used are Hatchbox PLA except for the silver which is 3D Solutech PLA.
Okay - that is progress and VERY good news! Fundamentally your printer is mechanically working fine, we have a software or firmware issue.
So I think I'm nearing the end of my ability to provide help, but personally I would try 2 things in this order: 1) de-install the slicer and delete any residual directories that may contain stored values and could possibly be picked up when you re-install it. I'm not sure of all the directory paths but I'm sure you can find them. Of course don't use any old gcode or yours, re-slice everything with default values.
If after all that the Prusa gcode still works but your newly sliced prints don't, the last thing I would try is a factory-reset of the printer. This erases everything that is stored in the printer. I have helped someone that had "tried everything" and it seemed to me like somehow there were calibration or other stale parameters that were not being erased and were causing problems or that there was corrupt data in the firmware, and after he did a factory reset (and the subsequent full calibration sequence) it magically started working. The way I look at it, after doing this you have a known good mechanical printer and a brand new, blank software load - there's absolutely nothing more you can do to strip the machine to it's bare minimum.
If there's anything else I can compare to my machine for you let me know.
Also FYI perhaps you jinxed me (LOL) but today I started printing the IKEA Lack parts in PETG which are available in gcode from Prusa so it must be good code. I've have printed PETG before a while ago, but it's been PLA for a while. I always have wondered what the inside of the nozzle looks like after switching between the two filaments as if there's any residual PETG in there after the switch to PLA it won't fully soften because of the lower PLA temps.
Note this is single filament - no wipe tower or filament changes.
Anyways, the first layer was "okay" but a couple of thin lines in the model did not fully adhere to the bed and some of the solid fill didn't look great and of course there was a bit of "stringyness" which PETG is known for. But it looked like it was probably good enough to get the second layer on.
THEN it happened! While I just happened to be watching closely (eye level with the build plate!) the extruder made some movements and I didn't see any filament come out at all! None! I thought maybe I was imagining it as it's black filament on a black bed and it's hard to see under the extruder. But for a few seconds nothing came out then a huge blob came out! It seemed like there was a blockage and then the dam broke and it all came out so I stopped the print. I've never had this happen before.
So as I sat there contemplating this and thinking perhaps the nozzle is somehow partially blocked I remembered the nozzle cleaner that I have and have never used. It looks like a steel hypodermic needle - one was included with the Mini but I also had bought a set years ago. So I removed the filament, heated the nozzle up to 250c, and when I inserted the needle it was actually quite stiff inside the nozzle. I worked the needle in and out, drawing filament out, cleaning the needle, working it back in and by the time I was done it was much easier to get the needle in and out.
So I started the print again and *AMAZING* the first layer looked spectacular, best first layer I've ever had with PETG. And also there was no stringyness at all, which I have always had. So I'm on the 4th layer of a very long print but so far it has made a significant difference. I wish I had done this sooner.
So I don't think it's related to your issue but I throw it out there as 1) for the first time ever I got a huge blob ejected from my nozzle and 2) if you switch between filaments, or perhaps even if you don't it's probably a good habit to poke up in there once in a while. I wonder if residual PETG from previous prints never cleared out, have "cooked" at a lower temperature for hours and hours and somehow gummed up the nozzle? I'd love to know if anyone has ever had this experience.
Figured I would tag this thread too. I just setup my MMU2 and the area where I print looks good, but my wipe tower messes up my prints which I thought was just me, but seems common. I was laughing over the fact that the area i want to print looks good and the waste area is actually breaking my prints of all things. I have been trying a simple test to print 5 fittings (one in each color), then create 5 other fittings as wipe objects with the idea to minimize waste. Well my little wipe block ends up being a magnet for globs of filament that the hotend picks up.
Is there a way to modify / configure that wipe block so its a different size or anything?
(The Black, White, Red, Yellow, Green fittings are the front pieces. The multi color ones in back are the wipe objects figured they might look neat)
The wipe tower size is calculated based on number of filament changes and other parameters (like wipe infill, etc); you certainly can't change the total size (that has to be fixed), but as far as changing the length and width I don't think so. (Of course you can change the location of the tower).
Back to basics - with the new MMU2S does it print normally single filament? Have you tried printing a stress-test in single filament mode to ensure everything is working okay with the current setup and filament? Are you running the latest firmware in both the MK3s and the MMU2s? And are you using the latest slicer with default settings? Your starting baseline should be with default Prusa slicer settings and the latest firmware and no mods that the MK3s prints perfectly. With that, you should download the Prusa gcode mentioned above as your first litmus test.
If so, read the above thread however some simple things to check: can you easily push the filament from the buffer, through the MMU2S and all the way own to the nozzle by hand? Are the MMU2S bondtech gears perfectly aligned with the filament? Are both the FINDA and the IR sensor working perfectly?
Lastly, in my post above I mentioned I got a blob for the first time printing PETG. Sticking the needle up in the heated nozzle seemed to help but it happened the very next time at the beginning of a print. It actually seemed like the PETG wasn't hot and fluid enough and that perhaps a cooler, more viscous blob inside the nozzle was impeding the filament and eventually the pressure built up enough and a huge blob squirted out. This was using Prusa gcode for an object I was printing and I used Prusament filament so I really doubted this was it, however moving the temp up 15 degrees (from 240, which is min Prusament PETG temp to 255) immediately allowed the filament to come out smoothly and the first layer (and subsequent layers) was perfect.
I mention this only because if nozzle temp was in fact my issue, perhaps it's also the cause somehow for folks having blobs on their wipe tower. A simple test might be to up the temp to the upper range for the filament you're using. (Also note different brands have very different ranges, e.g. Hatchbox is very different than Prusament for PLA).
with the new MMU2S does it print normally single filament? Yes
Have you tried printing a stress-test in single filament mode to ensure everything is working okay with the current setup and filament? Yes 17 hours
Are you running the latest firmware in both the MK3s and the MMU2s? Yes
And are you using the latest slicer with default settings? Yes
With that, you should download the Prusa gcode mentioned above as your first litmus test. I will try the gcode in the samples. I was using STLs from the samples and those failed too so I went to something smaller and faster instead of long (8-10 hr) prints for me to tune it.
Okay sounds like you have just about all the bases covered. I mention going right to the gcode as the slicer is yet another variable in the mix. I've worked with folks where the sliced models have issues but the gcode does not. (I recommend the "scorched earth" approach where you completely uninstall the slicer, delete any residual directories, and re-install). People forget they have tweaked a setting here or there or it gets picked up when they install a new version of the slicer.
The Prusa models are in 2,3,4 and 5 color versions - just start with the 2-color gcode in order to save time.
I loaded up the gcode and they are all PLA and I have PETG. If I modify the filament selection to use my filament is that a valid test still? I'm running Overture PETG.