[BUG] Extra extrusion after filament change #2502
Comments
|
This extra purge is a feature and the user shall count with that. It is there to prime the extruder and to reduce (avoid) the gap in the print caused by missing filament. Removing excess filament from the print is less a problem than adding back missing filament. |
|
Missing filament is bad for external perimeter but not for the rest (in most cases). But this excess filament is attached to the print always. So, at least, it would be nice to turn fan to 100% while extruding this excess filament not to attach to the print well. |
|
It was weird to understand this for me at first, but it works the following way: leave the excess attached to the nozzle while swapping the filament. As soon as the extruder starts moving, grab it with a quick motion to make a nice break at the nozzle. If you do that, the result is almost perfect (next to zero gap or blob). You cannot "grab" the filament unless you have some excess being extruded. |
|
I think this is related to the unfixed issue "M600 does unwanted extrude after 'is color clear?' confirmation" #197 I print almost only in PETG and this behaviour often makes the extra filament stick to the nozzle instead of the perimeter. The new lump on the nozzle is then dragged along making more harm than a tiny hole in the perimeter. I have found no way to turn this "feature" off... I understand that this extra purge is to build up the pressure in the nozzle - maybe there is another way to do that or reduce the problem:
|
|
Another relevant reference: "Blobs upon resume after M600" #904 |
|
The problem has already been described, it boils down to the fact that between "is color clear?" and the extruder moving some material has oozed already. We could retract just before the "is color clear?" prompt, but stuff like TPU would probably still leak a bit. LA cannot do magic either. It has to assume the nozzle is fully primed in order to work. Is it really such a big issue? When swapping filament some excess will always be there. I just pull it at the very end instead of doing it right after the prompt. I also print a ton of PETG, so I know really well the effect you describe. |
|
@wavexx |
|
The amount of extra extrusion is very much. It's like 10 cm with 0.4 nozzle. This can lead to collision during printing. Even with 1 mm nozzle it is more than enough. Moreover, only 0.6 mm nozzle is supported. |
|
This is definitely not a bug. I do remember instructions on changing filament, which explicitly says that you must grab extruded filament and hold it in place until hotend will move. Moving hotend will break filament from nozzle, leaving clean primed nozzle, and filament blob on your tweezers. |
|
"I do remember instructions on changing filament, which explicitly says that you must grab extruded filament and hold it in place until hotend will move. Moving hotend will break filament from nozzle, leaving clean primed nozzle, and filament blob on your tweezers." This doesn't work very well for me. When I grab hold of the filament change blob if often breaks loose from the nozzle. How does this help with priming and pressurizing? The extra extrusion is into open air just like the filament change extrusion 2 seconds earlier. The only need for this is if you wait a long time after the filament change to continue printing. How about a timer? Less than a few seconds - no extrusion; Longer - do the extra extrusion. |
|
Sorry I posted #2801 as it is a double post. However, I will add that after the filament change is done, and while the printer is moving to the start position, the extruder is turning, so it really is almost impossible to prevent a blob, especially with PETG. |
|
Forgot about this until now.. I had a few filament swaps yesterday, and I might have to agree. As @sledjunk said the extrusion keeps going until XY are in-line with the print. Sometimes grabbing the filament as late as possible is not always easy on large prints. Thinking again, I'd still do an extrusion before starting to move (the filament just oozes out even if you're fast with the button), followed by a quick retract and then move. This way if you grab the last bit of filament before (or during) the extruder moves there's less ooze going on. I tried this by simulating a change with some gcode and seems to work "good enough", certainly not worse than before even if you click "ok" after ~30s or so. Some ooze is still present even with a retract though. The move back to the print is slowish.. I'm not sure if this speed is fully intentional? |
|
I tried again yesterday and I used a strip of plastic (actually a ruler that came with overture PETG) and held it to the nozzle when it landed on the job so that it caught most of the blob. It helped somewhat, but you almost have to ruin one print so that you know where the nozzle is going to land on the job. |
|
If there was a beep and 1 second pause just before moving back to print you could easily grab it. This would be good at the start of prints too. I've gotten used to grabbing the ooze just before a print starts by watching the temps and knowing when it's about to start. If I don't grab the ooze, it's hit and miss if it will fall off during leveling and be laying in the middle of the print area. |
|
I haven't really had an issue with start of print because of the purge done after bed leveling and before the print begins |
|
I don't have a problem after the purge starts, but once in a while the preheat ooze will fall off while leveling and then get into the print. It happens often enough that I got in the habit of grabbing in just before the leveling starts. |
|
Yea, this is still an issue even on the latest version firmware and PrusaSlicer. Ive gotten pretty good at grabbing that extra filament that leaks out but it would be nice if I didnt have to babysit the change over so much. |
|
I have this issue repeatedly on both my MK3S when using PLA. I use an M600 Tool Change Command during the same layer to get multiple colours for signs and plaques. This extra purge is un-necessary as you have just wasted a load of filament flushing the old colour and making sure its ready. It just creates blobs sufficient that the nozzle crashes on the print and I've had it wreck production pieces over and over again. Its not oozing, its not the wrong filament change technique (I do 20 changes on some prints and always grab the excess and pull it away at the very last second.). It just sits there and blows a blob of filament as if its trying to wipe. |
|
I remember some time ago, when I did some prints with many color changes per layer (over 10), the only reliable way to do the filament change was to use at the slicer a fine tuned purge tower, even without the MMU. |
Thanks, actually I always turn off the purge tower but now I can see the use of it for fixing this issue. Thanks Panay! |
I tried this and it worked perfectly! No sign of any blobs! Thanks very much for the help. |
|
Summary of the issue The problem: unwanted blobs/strings of filament ruin prints for users because there is very little time to remove the extra extruded string and it is surprising and undocumented. Surprising: No extra extrusion into the air happens when loading filament at the start of a single color print. There are two suboptimal workarounds. Workaround 2: Manually create a purge tower object in PrusaSlicer. Right click bed > "Add shape" > rectangle > size it about 3 cm long, 3 mm wide and height above the final color change layer. Position close to the south-east corner of the build plate. In the right pane list of objects drag the rectangle to the bottom to make it print first at each layer. Press "Slice now" and use the g-code preview sliders to double-check that print starts at the waste tower after color change. That's suboptimal because print time increases and the manual steps take time. The problem has been reported multiple times for at least 4 years. 2017 prusa forum and issues years 2017, 2018, 2020. New users still run into the problem - see comments on Color change help page. Things that could fix or mitigate the problem
|
|
Any chance this is going to ever get fixed? If an extra purge extrusion is desired, users can add that in. We need a way to not have it push extra filament when it doesn't need to. Like in comment #2502 (comment), I have a single extruder and do color changes and after the color change have to babysit the purge. Sometimes there's room to do Workaround 1 listed above, but not if I'm printing something that goes all the way to the resting edge where the hotend moves for the change. |
|
I also experience this hm |
|
Hello! Thank you for the feedback on this. Unfortunately, no timeframe yet or confirmation of a possible fix at this moment. In a way that extra extrusion does ensure the correct pressure and flow are there. Did the workarounds proposed on previous comments improve the situation so far? Even just being ready with pliers to remove that excess can do, in addition to all the workarounds on the previous comments. Alessandro Pantaleo |
|
This is already a great improvement. |
|
A partially related issue where the nozzle moves back to the last position on the print - and releases some different color oozed filament in the wrong location - will be addressed via Prusa Slicer (issue prusa3d/PrusaSlicer#2672, PR prusa3d/PrusaSlicer#9036, commit prusa3d/PrusaSlicer@540b800) . The specific issue discussed in this thread instead - about the extra extrusion after filament change - should be solved with PR #4534. However please notice, even if merged, this feature is not enabled by default because experimental. Please go through the Pull Requst to learn more and the feature should be available in the next firmware release (but not active by default). Michele Moramarco |
|
Closing as no interaction |
|
Please reopen this issue, it is not completed. The issue has been reported since 2017, so around 7 years now. Do you really need us to post a new comment saying the issue still exists every 30 days? |
@Prusa-Support no it is not explained in those comments. The thing that is unexplained is why this extra (surprising and unwanted) extrusion of filament happens when manually changing filament color during a print after then new filament is loaded and some of it has already been extruded and directly after the user presses the button to resume the print with the new color. After that button press it takes around 1 second for the printer to move from the new filament loading position near the front right corner to whatever location the to-be-resumed print is at. Why is it according to you necessary to extrude extra filament in the air during that move? If your concern is that after loading the new filament the user waited too long before pressing the button to resume and that causes some build up of heated filament that has to be disposed to avoid a blob on the real print then a proper fix would be to keep track of the time waiting for the user button press and, if too long, make the button press trigger some extrusion without moving and show a new button press prompt. Note also the difference between this extra extrusion and what happens when starting a regular print: on a regular print start, after extruding a line of filament near the front left edge of the plate the head moves to the start print location and only then extrudes filament. It does not extrude any extra filament in the air between the front left edge and the start print location. For some reason that is not necessary - but a very similar extra air extrusion is claimed to be necessary after a filament color change? Finally let me ask Michele Moramarco and other prusa support people who comment in this thread, have you yourself actually reproduced the issue on a printer and seen it with your own eyes? |
I STRONGLY ENDORSE THIS COMMENT. Why do the devs REFUSE to fix this despite numerous complaints and multiple proposed fixes ranging from simple to complex? PLEASE FIX THIS AND PLEASE STOP TRYING TO CLOSE THIS BUG EVERY 30 DAYS!!!!!!!!!!! |
|
BTW, the claim that the extra extrusion is "needed" for filament changes in case the user stepped away from the printer and the filament needs to be pressurized is BOGUS or at least is inconsistent with how Pauses are handled. So, I call BS. Specifically, when a Pause is triggered followed by a Resume, there is no similar "extra extrusion" of filament -- even though it is very common that a person may fail to immediately respond to a Pause, potentially causing real depressurization of the extruder. |
|
@floslo @puterboy Thanks for your feedback. Did you read #2502 (comment) ? Also please check my comment in the PR #4534 (comment) You may think of that extra extrusion as a bug and annoying but most are okay with it and "know" to grab the extra extrusion before it moves back to the print to ensure the print continues without any gaps. So feel free to activate You need to understand that changes like this have a big impact as the ooze prevention may work for a setup of material, nozzle size, etc. but needs different settings for other combinations. We will not change it for the time being in the stock firmware. |
|
@puterboy Please stay polite
That kind of comments do not help at all and I tend to step away from these kind of issue tickets. |
|
@3d-gussner - You have got it wrong. This is not about before the move back to the print, it's AFTER the move. The nozzle comes down onto the print to continue printing and blows a bubble of filament unnecessarily before it continues printing creating a crash risk / print artifact. It does not do this when a purge tower is deployed. Based on your response, you have never tested this to see what we are referring. |
|
@ChrisWhite1985 I never had an issue during manual color changes and never saw the "blows a bubble of filament" on my prints. |
|
@3d-gussner Make sure this is during a single layer colour change. I.e. 2 colours on same layer with manual filament change M600. |
|
Here my test scenario:
My conclusion:
I will try to prepare a animated gif.
|
|
As the MK404 SIM doesn't simulate nozzle pressure I will try that on the real hardware later. |
|
@ChrisWhite1985 I think this github issue has comments talking about 2 or more separate problems. My comments are about the extrusion-immediately-after-button issue: ~5mm extrusion occurs immediately after the user presses the button to confirm that the filament changed correctly, at the same point in time as the head starts to move from the front right position. Your issue could perhaps be called the extrusion-at-print-object issue: it occurs later, when the head has already moved back to the last print object position and is to start extruding filament onto the print object. (Apologies if I'm misdescribing/misunderstanding here, I have not myself experienced that issue.) |
|
@3d-gussner thanks for reopening and testing!
How do you know most are ok with it? How do you know most know the practical steps to deal with it? As I noted upthread #2502 (comment) the issue is surprising because it is not documented. So how would someone who for the first time attemps a mid print filament color change know that the issue will happen or how to deal with it? Let me repeat two earlier suggestions on what Prusa can do to mitigate
The most popular pratical method to handle the issue from this thread can be summarized as: 1 Let the first, automatically extruded larger amount of filament hang between the nozzle and the bed to which I add 4 .... and hope that as the head moves that extra 5mm will break off clean from the nozzle and stick to the filament string you're holding down. If it doesn't then react with lightning speed and move your hands to grab a piece of paper and wipe the dangling filament residue from the heated nozzle tip during the ~1s it is moving from the front right back to the print object. Make sure you don't burn your hand on the hot nozzle during this very quick hand movement. Also make sure you don't accidentally bump you hand into the print job. If you fail your print may look ugly or be ruined. Because in my experience sometimes that ~5mm will stick to the nozzle. Especially when printing PETG that extra filament can twirl or bend back up and stick. Lastly apologies if I'm missing something but can you explain why that extra ~5mm is necessary in this situation but not at the start of a regular print (right after the printer has placed a line of filament at the front left edge of the plate and while it is moving to the print start location)? I'm asking because the time it takes for the head to move is roughly the same in both those situations. If that difference can be explained in a short but clear way then consider also adding that explanation to the help page, so users can learn. |
|
@floslo Thanks for the suggestion, I will forward it to the content team.
I am not the Prusa Slicer settings Pro but here my thoughts:
Take some pliers to hold the filament 😉 |
|
@3d-gussner thanks Pliers vs finger: For me using a finger (in glove if the user feels the plate heat is uncomfortable) is less error prone. With pliers I've accidentally moved the filament as I grabbed it, causing it to break off or weaken at the nozzle before the button press, which in turn caused the extra 5mm to not stick to the rest of the grabbed filament. That the success of this user step is sensitive to very small hand movements underscores that the best solution would be to find a way to get rid of this step.
Ok, that does make sense. If that is indeed the difference then it also suggests another possible solution to consider: in jobs where the user only changes color once and where there are no print objects near the front right part of the plate then instead of the extra 5mm extrusion in air make the printer extrude a line on the plate just like when starting a regular print, only now along the front right side of the plate. PrusaSlicer could be made to detect if a job is like that. My guess is that would cover a majority of color change jobs. For example a user printing some object near the center of the plate and wants embossed text on top in a contrast color. |
|
Alternatively, if the current layer has enough infilled area available, could the extra 5mm extrusion perhaps be made into the infill area? There'd be some structure present to conveniently "wipe" the nozzle against, and it would be invisible since it's infill. |
I had the same idea to define per sliced layer a safe ooze location in the empty space of the infill, but I got few comments on that:
|
|
I’ve been closely following this thread and it appears to me that the issue of extra extrusion after a filament change remains unresolved in the firmware, even after 4+ years. It seems that the complexity of handling this correctly might exceed the capabilities at the firmware level and perhaps should be delegated to the slicer, which has more context about the print. I propose enhancing the flexibility of the M600 firmware command. We could introduce a parameter to specify the amount of extra filament to extrude. This way, users can disable extra extrusion by setting the length to 0 or specify a suitable length for the current printing context (both filament and 3D model). If the parameter is omitted, the behavior would remain as it is today, ensuring backward compatibility. This change would also empower the slicer with a more versatile M600 command that can be used in different ways, depending on the printing context. The slicer could vary the extra extrude depending on the 3D model or filament type, or even disable extra extrude and add a G1 Exxx to simulate pressure build where needed. It opens up a multitude of solutions. The introduction of the M600 parameter also eliminates the need for custom firmware builds, which can be challenging for non-expert users. I believe this approach could provide a more flexible and user-friendly solution to this long-standing issue. I eagerly look forward to hearing your thoughts on this suggestion. Thank you. |
|
Its pretty on par for Prusa to keep selling "upgrades" every few months but fixing a simple yet severe issue with filament change remains ignored. I guess the official "FIX" is grab it real fast while its moving and don't burn yourself or miss, or else you just wasted 6 hours of print because sending the extruder back the front to wipe again is too much after over 4 years of asking |
|
Exactly! Prusa seems to have really dropped the ball lately on hardware, software and pricing. Sadly, I don't think my next machine will be a Prusa -- I really wanted to continue supporting this company... but they just need to keep up with the times... |
i3 MK3S FW 3.8.1
Should not extrude extra filament but do retraction. Since there's no default retraction length, there should be Marlin instruction added which defines default retraction length. There could be retraction instruction as well.
The text was updated successfully, but these errors were encountered: