Ultimaker2 Feeder System - Improvements and Ideas

[shurik 87]

Didier Klein - which version of the feeder do you have? Robert put several of them.

 

[DidierKlein 729]

I printed this one:

https://www.youmagine.com/designs/alternative-um2-feeder-version-two#

 

[Blizz 53]

From friction point of view, I don't know if straightening the filament actually helps

 

3Poro: considering the bowden tube I'm sure it won't for all the points you mentioned.

But the subject was the wear & tear on the bottom of the feeder plastic. I'm pretty convinced it will help there, the bowden does a pretty good job of keeping the filament in check imo.

 

[Blizz 53]

Ok after some test i have found 365 Steps to be very good usable.

to apply them copy this to layer 0 on any gcode:

M92 E365

M500

 

@Takei: So is this the only thing we need to change then? One time in a print? Or is the retraction update etc still needed? There are so much different things going on in here it's hard to follow

 

[takei-naodar 0]

The M500 command saves the change, so you will only need to do this once and you will not need to change any other settings. Although i am not sure if this setting is kept after a firmware update.

 

[Blizz 53]

@Takei: Most important for me is the only once part. I guess that I'll add the commands again to the first part I print after an update to be sure.

I asked because the first print I did with the MK8 mounted had retraction issues. It left strings all over the place. But then I reset the firmware back to factory settings (couldn't find the default E-steps value) and remounted the original gear and it still had the issue. So was related to other things

 

[eldrick 65]

The original E-steps value is 282, at least on mine purchased in April.

 

[Blizz 53]

The original E-steps value is 288, at least on mine purchased in April.

 

I think I saw 282 mentioned somewhere.

Anyhow: I set the 365 that was calculated higher up for that MK8 gear. Unfortunately I have to put the flow rate on 1.05 to not get any underextrusion. So is my hardware different somehow or was that e-step value wrongly calculated?

 

[mnis 11]

In the Cura Machine settings e-steps can be entered under Ulti-GCode Flavor. The parameter 288 is there but apparently too high. With the original knurled wheel a functioning value is closer to 144, I think.

Currently I use an alternative engine with half the resolution, but more power, the normal knurled wheel, use the normal RepRap GCode Flavor with 165 to 185 non-cubic e-steps. This works very well. Also the RepRap Flavor offers additional setting options.

No matter, I would always print a small test cube. With each new cube would I change the e-Steps in five / ten steps. On closer inspection you can see and feel how the surfaces change significantly. If the surfaces feel haptically smooth, then it need only minor fine tuning.

Markus

 

[sebastian 0]

Hi guys,

just uploaded the feeder version that we are using on multiple UM2 here to YouMagine:

https://www.youmagine.com/designs/yet-another-ultimaker-2-feeder

Basically, this is a combination of Roberts and Geeks feeder. Pretty happy with that one!

Sebastian

 

[Blizz 53]

Looks pretty awesome. Might replace the feeder body with that one, will allow me to get rid of the filament guide under the power cable.

 

[IRobertI 520]

That's a pretty sweet mash-up Sebastian

I finally sat down to do an update on my version and if anyone feels bored enough to try it out before I change the files on Youmagine I'd appreciate it. I'm mostly interested to see if the screws bottom out and lock in place while still allowing the latch and arm to move. There's a tiny margin of error there and I'm not sure all machines are identical enough for this approach to work.

You probably can't really tell a difference in how it looks but I made a couple of tweaks. I got rid of the spacers that everyone struggles with. I made the latch slightly beefier. The arm is now printed lying down to improve strength and I rotated the nut-hole to make the wall ever so slightly thicker. The hole in the arm was reduced to reduce wobble. The snap-on guide was beefed up slightly for strength and I also tightened up the tolerances a bit. The yoke is slightly higher to reduce play and I made it a bit shorter again.

So, mostly incremental changes, nothing revolutionary.

(If you already have a copy of my feeder you don't need to re-print the hub adapters, they are unchanged)

https://www.dropbox.com/s/y4hfg6sf2mhsmbm/feeder_v5_noSpacers.zip

 

[Blizz 53]

Would be kind of cool if you could combine your changes with the filament guide, I was just considering changing the feeder body by that one. I'd really like to get rid of the "filament below the bottom of the machine"-thing.

I've reprinted the v4 of your feeder btw and haven't had problems with the arm since. I do have to admit that when opening it I make sure I apply pressure close to the yoke. When it broke last time I was pushing on the end near the latch and that was probably part of why it broke.

I'll try to print out v5 and see what gives. I did mess up one of the body screws last time so I don't know if I can still get it off Love the fact that you got rid of the spacers, my god those were annoying

 

[IRobertI 520]

I'm using the normal spool holder and haven't had any problems, I haven't had the grinding at the side of the hole either. Maybe people aren't putting in the little bowden piece? I dunno. But I'll consider it.

The arm should be much stronger when printing in this orientation so now it shouldn't matter where you push (I've always pushed near the top though without breaking the 4-5 versions I've had on there).

Yeah... it seems those things really messed with peoples heads hehe.

 

[Blizz 53]

To be honest: I didn't put in the piece of tube. I will for this one. It looks a lot more solid.

If you don't have issues with the hole eroding then I'll get rid of the extra guide.

 

[Blizz 53]

Arm is printing right now.

Minor detail regarding the body: I've had to rotate it 180º. The wider part for the snap-on is at the bottom in default orientation so most of the piece is floating in the air and will need support like that

Will keep you appraised of my progress.

 

[IRobertI 520]

Dammit, I thought I fixed that... Updated the zip-file so it should be right now.

 

[Blizz 53]

I printed it, messed around with getting the support out and mounted it.

 

No real issues found.

 

Printed "the tube" and experienced some underextrusion at 10mm3/s but then I recalled I forgot to up the temperature, so all good!

 

[IRobertI 520]

Nice, thanks for helping me A couple of questions. How was the hole size on the latch and arm, any problem getting the screw through? (I see you're not using the stock screws there but the diameter should be the same) And the support, how much of a pain was it to remove? I considered leaving the top of the arm exposed to get rid of the internal support there but I like the closed look better.

 

[Nicolinux 288]

Hey Robert,

Regarding support. If you use Simplify3D you could prepare/slice the files for others and add support where you think it fits. Another bonus of you slicing the stuff is that you built it and would know how strong (how much infill) the parts should be.

 

[IRobertI 520]

I added in manual support into the STL already. Maybe I should've mentioned that so people don't enable it in cura themselves... oops.

As for slicing and providing gcode. I'm a bit torn on that. Some people might use those files, other might not because they want some special setting. So in the end there isn't much control either way.

 

[Blizz 53]

How was the hole size on the latch and arm, any problem getting the screw through?

Yes because I kind of messed up the stock screws. But these are also 3mm and although I had to "screw" them all the way through (couldn't just push them through): a lot less effort than the spacers. I have a bit more resistance now when opening the arm, but nothing to bad. I don't overdo the tension on the screws.

 

And the support, how much of a pain was it to remove?

 

Actually looking back: Not too bad. It's just that with the changes I wasn't fully sure if I had to leave the support for the holes or not. They looked pretty solidly attached and since no more spacers were required I was afraid they would be needed initially. So it took a bit more effort to leave them alone. After I printed the body I figured out quickly that they needed to go. If I had known that earlier it would've be less effort

I would just leave the STL files and don't pre slice it. You don't have "special requirements" in this one like the spacers before and not everyone wants to print these at 100% infill (or some people actually do). It's hard to please everyone.

 

[Blizz 53]

Was printing the snapon and didn't look at it. At the end it turns out that Cura decided to close the hole lol. Pretty useless like that

Nvmd, too quickly. It just started printing it when I aborted. Too tired, gonna try again tomorrow

 

[BasVanDeursen 0]

Hi everybody,

It's been a bit silent from my side lately (it's quite busy here), but I would like to let you guys know that I am still following this thread. I've also tried out Robert's feeder design on one of the UM2 printers here at the R&D department. One of the things that I particularly like is the quick release system. Besides that I think that it is good that the entrance of the Bowden tube is really close to the feeder wheel (for printing with flexible materials). I've also implemented this in the latest revision of the UM2 feeder housing.

Unfortunately for the UM2 we are going to have to act within the boundaries of the current tooling (injection molding dies are quite expensive tools). However I'll definitely keep all the good ideas that I see on this forum in mind for a future feeder design.

The new machines are now fitted with a revised version of the feeder housing which includes the metal grommet at the feeder entrance. Besides that we are currently testing with a slightly adapted feeder lever in order to provide better guidance for the filament during filament loading.

In the future I would like to have some more torque on the feeder axis. At the moment we are testing with different (geared) motors to obtain a higher torque on the feeder axis. Besides that we have performed some research in order to get more insight into the resistance induced by the various elements in the extrusion train. I've written a test report about the method that we used and about the first results. If anyone would like to read this report: let me know and I'll make sure that you'll receive a copy.

Keep up the good work and have fun printing!

-Bas

 

[bertho 6]

Thanks Bas,

I am interested to read your report.

I am monitoring the actual injection force on my UM1 which is very helpful to be able to see any how the system works and how much safety margin there is.

Bertho