Selector B18 - Filament hitting a wall

[corporategoth]

I've printed these parts with a Prusa MK4, in Prusament PET-G, 0.15mm Structural.

If you have a filament that likes to 'curl' or as a natural curve, then often it will 'follow' the ball baring in the selector. The hole for the filament to exit the selector is too small, or at least, the guide for filament to follow the path is not big enough. When this happens, the filament hits the wall inside the selector, and cannot exit the selector.

The selector should be designed to either better constrain the filament so that it can't curve (difficult since that will hinder the ability to move the ball baring), or the channel to guide filament into the exit hole needs to be bigger to allow for filament that likes to curl/curve upwards.

I had this problem happen repeatedly with my Polymaker PET-G filament.

[davidkerny]

Hi!

The filament curling, as you describe, usually happens only on the first load of a new filament. I wasnt able to reproduce it with a filament that has a tip formed in the hotend. That would mean, that straightening the filament end (as the MMU3 handbook and assembly guide suggest doing before preloading a filament), would fix it.

The selector B11 (to be used in combination with blade holder B14) had a larger filament entry after FINDA ball, as there is on the original MMU3 selector. This larger entry point used causing three things:

1. curled filament was able to load more easily. 2. Unloading a depleted filament with a curl at the end was often unsuccessful. (spooljoin) 3. strings were entering the filament entry often jamming the filament transport path. The smaller entry point has fixed all these issues. Not only that, Now, I was able to print very stringy materials unlike before, as the filament strings were diverted sideways, and exited the finda ball chamber more easily after fine-tuning the geometry.

Due to the larger side openings around the FINDA ball, the conical entry cannot occupy the same space as it used to on the selector B11. I tried reconstructing the shape to fit the most up-to-date selector. The result are some geometrical weirdnesses but it should work similarly to the B11 version. Please, test it to see if it helped with the curled filament:

https://github.com/davidkerny/BUZA-ULTIMULTI/blob/main/OLD-OTHER/SELECTOR_B19_EXPERIMENTAL.stl

[corporategoth]

I was thinking the entrance could be more keyhole shaped (with the sides being angled like they are now to guide towards the hole), something like:

The blue shape. That would catch anything that went high and guide it towards the hole. That would not affect the ball as it's much wider than the width of filament.

However, I've not tested this new part yet. I'll print it up and see how it goes.

I had not seen that about straightening the filament before first loading - but I'm more worried about filament swaps - sure, I could straighten it on first load, but if the filament has a tendency to curl (because it's stored in a circular spool), then after printing with it for a while, then unloading, then re-loading it for another layer will mean it's likely gone back to it's curled state. So something needs to be done about it I think.

And sure, it might just be a problem with some specific filaments. But not all my filament is Prusament :D

Thanks :)

Preston

[corporategoth]

Here is what I managed to create using PrusaSlicer to add negative volume modifiers

Not tried it out at all yet though. I'll test the experimental B19 you posted first. So far it's not failed to load/unload filament. But if it does start failing, I'll test the part I modified.

[corporategoth]

@davidkerny FYI - your B19 seems to be doing the trick, I am not hitting the wall anymore on my MMU3 print.

I didn't even have to use my modification to it (with the longer channel) - your printed part was sufficient.

Thanks for the quick update!

[corporategoth]

One other thing I will mention is that I think the holes for the brass tubes are slightly too big. They don't "hold" onto the brass tubes well enough so the brass tubes will slip out of the selector, and end up in a random position on the rods they are sliding over, as opposed to sticking to the selector.

[firetech]

Just to add my 2 cents to the discussion:

I've been running the UltiMulti as intended (i.e. with Selector B18) for some time, without any issues, mainly printing PETG. When I loaded some PLA that was nearing the end of its roll yesterday, I experienced this issue. It was happening repeatedly, basically every time the filament was loaded again. Switching to the B19 selector seems to have solved it, even if I intentionally make the filament even more curved.

Furthermore, I don't have any issues with the brass bushings sliding out, I'm not even sure I can get the ones in my B18 selector out without damaging anything (luckily, I had extras)...