MMU2 - another solution for layer skipping due to filament length
So... My experience with MMU2 was fine, but I kept getting missing layers, even after 1.0.1 upgrade.
When I took a closer look, i noticed that the filament did not reach the bondtech gears because the tip does not have a constant shape, hence, it did not have a constant length. The issue seemed even more difficult that when I played with the settings I used to get different tip length.
I could go for a long session of calibrating but that seemed like the long way and I am lazy. Instead, I added a couple of springs (taken from a ball pen) to the MMU2 outlet, and used longer screws + extended the bowden length calibration at about 1-2mm for each filament.
Now, when the filament reaches the bondtech gears, it pushes the MMU outlet a little, so the filament length can vary to some extent.
The MMU outlet is moving up and down on a smooth area of the screw to avoid grinding of the plastic part.
Maybe it can help someone who has the same issue...
Its the same tolerance compensation mechanism you can find in your extruder. The only difference is that here the springs are compensating for length tolerance, where in the extruder they compensate for thickness(diameter) tolerance.
Re: MMU2 - another solution for layer skipping due to filament length
I'd had a similar idea but thought of designing an additional attachment with a spring.... Yours is a nice elegant solution, I might try it myself.
Re: MMU2 - another solution for layer skipping due to filament length
Nice solution 😀 and good thinking... will remember that when I get my MMU 2.
Have a look at my models on Printables.com 😉
Re: MMU2 - another solution for layer skipping due to filament length
I could go for a long session of calibrating but that seemed like the long way and I am lazy. Instead, I added a couple of springs (taken from a ball pen) to the MMU2 outlet, and used longer screws + extended the bowden length calibration at about 1-2mm for each filament.
That's brilliant, and wonderfully simple.
It's clear that the path from the MMU to the exruder isn't perfectly consistent (you can see this in the hysteresis that's visible when you calibrate the bowden length), but the design of the system assumes that it will be consistent.
Allowing for a few millimeters of sloppiness seems like the right way to go--a lot easier than trying to make the rest of the system more rigid. And, since the only function of the MMU is to get the filament to and from the bondtech gears, we don't really care if it's a little sloppy about how it does that.
Re: MMU2 - another solution for layer skipping due to filament length
Damn that's really smart ! I will add this !
Thanks for sharing Newguy !
Re: MMU2 - another solution for layer skipping due to filament length
I am not a smart man.
Why does this help at all?
Is the theory that the MMU gears would try to feed filament into the bowden tube, but meet resistance at the entrance hole (due to tip length/size), then fail to push all the way to the extruder gears?
Re: MMU2 - another solution for layer skipping due to filament length
I am not a smart man.
Why does this help at all?
The idea is that this lets you safely calibrate the bowden length on the long side, so the MMU loads a little too much filament but it's OK because the springs take up the excess.
A fairly high percentage of load failures happen when the filament almost loads all the way but doesn't quite reach the bondtech gears for one reason or another. This might help with that.
Re: MMU2 - another solution for layer skipping due to filament length
That's really clever!
I've mostly done prints under a couple hundred layers so I've kept my changes to a relatively modest number - which I've been able to make work through other tuning, but I like what you have!!
Have you played with extra restart distance, or you just calibrate for longer tubes?
I maintain an informal list of San Diego, CA 3D printing enthusiasts. PM me for details. If you include a contact email and I can add you to the informal mailing list.
!! Warning!! Re: MMU2 - another solution for layer skipping due to filament length
I just tried this, and you need to be very careful about how far you screw in the screws.
It's possible, even easy, to destroy the FINDA probe by screwing in the top screw too far. With a longer screw there's nothing preventing the screw from ramming right into the probe.
Re: MMU2 - another solution for layer skipping due to filament length
Good point - the top screw can touch the FINDA sensor if you use a long screw (didn't quite see that one before the post above me), I simple used a screw I had in hand that fit), but you can easily cut the access threads of one of the screws from the spares, having about 1cm of threaded portion of the screw and make sure the top part (coser to the Allen) is flat to avoid grinding of the plastic.
Hope no FINDA was damaged 😕 😕
Re: MMU2 - another solution for layer skipping due to filament length
Clever!
Re: MMU2 - another solution for layer skipping due to filament length
If I ever receive mine I'll be sure to keep that in mind. Still don't get why they don;t use the built in cutter to ensure a consistent tip, it's there after all. But instead they apparently opted to disable it?
XL (5T) / MK4 / MK3S+-MMU3 / MK3S / MK2.5S-MMU2S / Mini / SL1S
Re: !! Warning!! Re: MMU2 - another solution for layer skipping due to filament length
I just tried this, and you need to be very careful about how far you screw in the screws.
It's possible, even easy, to destroy the FINDA probe by screwing in the top screw too far. With a longer screw there's nothing preventing the screw from ramming right into the probe.
From the initial picture, it looks like a M3x20 screw would be a good option, is that too long for this (assuming you don't torque the &%*! out of the spring)?
Re: !! Warning!! Re: MMU2 - another solution for layer skipping due to filament length
I just tried this, and you need to be very careful about how far you screw in the screws.
It's possible, even easy, to destroy the FINDA probe by screwing in the top screw too far. With a longer screw there's nothing preventing the screw from ramming right into the probe.
From the initial picture, it looks like a M3x20 screw would be a good option, is that too long for this (assuming you don't torque the &%*! out of the spring)?
The problem is that the screw hole goes all the way through to the cavity that holds the FINDA. The only thing that stops the 10mm screw from hitting the FINDA is that it isn't long enough.
If you use a longer screw, there will be no torque on it at all unless you're actually hitting the probe. That's what makes it so easy to damage the probe.
Re: MMU2 - another solution for layer skipping due to filament length
Not having a MMU2 - I have a basic question. If the amount of filament extruded is a little long, why does that matter? Doesn't the extruder gear start turning a bit before the filament arrives, and hence it shouldn't grind? This just seems like it isn't rocket science. Is the design one that has the MMU feed the filament the calibrated length, stop, then enable the bondtech extruder gear? If that is the case... it seems like a lousy way to do it.
Re: MMU2 - another solution for layer skipping due to filament length
Not having a MMU2 - I have a basic question. If the amount of filament extruded is a little long, why does that matter? Doesn't the extruder gear start turning a bit before the filament arrives, and hence it shouldn't grind? This just seems like it isn't rocket science. Is the design one that has the MMU feed the filament the calibrated length, stop, then enable the bondtech extruder gear? If that is the case... it seems like a lousy way to do it.
After experimenting with it and on the advice of others in the forum, the real issue is that the default setting for the bondtech speed when the filament hits the bondtech is too slow.
This is a simple thing to fix, and I'm surprised Prusa hasn't pushed out updated filament profiles yet.
Under "Advanced Settings" for the filament profile, change "Loading speed at the start" to somewhere around 19. 19 works best for me, but your mileage may vary given the inconsistent MMU gear diameters.
Re: MMU2 - another solution for layer skipping due to filament length
Not having a MMU2 - I have a basic question. If the amount of filament extruded is a little long, why does that matter? Doesn't the extruder gear start turning a bit before the filament arrives, and hence it shouldn't grind? This just seems like it isn't rocket science. Is the design one that has the MMU feed the filament the calibrated length, stop, then enable the bondtech extruder gear? If that is the case... it seems like a lousy way to do it.
The issue is that the MMU feeds the filament much faster than the bondtech gears so the filament hits the gears and it binds causing the MMU to grind away the filament. You can perform an individual filament calibration and calibrate the feed distance on the MMU but it's incredibly inaccurate.
For example, during calibrating, I'll hit the feed button and the filament will feed perfectly. I'll hit the retract and then feed it again. On the second feed it will feed it too far. I'll adjust it again and the next time it won't feed enough. It's never consistent.
Re: MMU2 - another solution for layer skipping due to filament length
Not having a MMU2 - I have a basic question. If the amount of filament extruded is a little long, why does that matter? Doesn't the extruder gear start turning a bit before the filament arrives, and hence it shouldn't grind? This just seems like it isn't rocket science. Is the design one that has the MMU feed the filament the calibrated length, stop, then enable the bondtech extruder gear? If that is the case... it seems like a lousy way to do it.
The issue is that the MMU feeds the filament much faster than the bondtech gears so the filament hits the gears and it binds causing the MMU to grind away the filament. You can perform an individual filament calibration and calibrate the feed distance on the MMU but it's incredibly inaccurate.
For example, during calibrating, I'll hit the feed button and the filament will feed perfectly. I'll hit the retract and then feed it again. On the second feed it will feed it too far. I'll adjust it again and the next time it won't feed enough. It's never consistent.
I haven't had that happen to me. When I have caught it skipping layers, you couldn't even see the filament because it didn't push it close enough to the gears. Usually happened after one of the filament jammed and had to be manually fixed.
Re: MMU2 - another solution for layer skipping due to filament length
I haven't had that happen to me. When I have caught it skipping layers, you couldn't even see the filament because it didn't push it close enough to the gears. Usually happened after one of the filament jammed and had to be manually fixed.
I have that happen too. Every one of my prints fail by either having a missed layer or the MMU strips the filament.
Re: MMU2 - another solution for layer skipping due to filament length
I could go for a long session of calibrating but that seemed like the long way and I am lazy. Instead, I added a couple of springs (taken from a ball pen) to the MMU2 outlet, and used longer screws + extended the bowden length calibration at about 1-2mm for each filament.
Thanks very much for sharing ! Good work -and good thinking 😉
I have that happen too. Every one of my prints fail by either having a missed layer or the MMU strips the filament.
Lot of times for me too. Was actually thinking to find a solution for exactly this when I found this thread. Trying yet another print, if it wont work either, the MMU will be placed very close to the bin. 😡
Mk3S+, Mk4 MMU2S never worked properly