BasVanDeursen

Cool idea!

I asked Erik yesterday and he said he used the 400 motor. -Bas

Hi Everybody! We are indeed working together with Mutley3D to make an upgrade kit that would fit the UM2. This kit would be aimed at people who are not afraid to do some tinkering on their machines. However, at the moment we are still in the prototyping phase, so it might take a while before this kit will become available. So "should be available soon" might be a bit too optimistic. Kind regards! Bas

As far as I know all UM2's have 400 steps/rev motors. It might be that Erik did not use his stock UM2 feeder motor for his Flex3Drive setup. I should ask him about that next time I see him. Another comment that Erik made to me the other day: make sure that your retraction length is set to 1.5mm or something in that range. The drive wheel is so close to the nozzle that you won't have to compensate for any slack in the Bowden tube so you can use a low setting on your retraction length. -Bas

But you are building it onto a UM2?

I did not do an awful lot of printing with the Flex3Drive yet but during the first bunch of prints I noticed that there was some stick-slip between the gears. I put some grease on the gears and that solved the problem. -Bas

I can't find them but a 200 steps/rev stepper should do the trick. I think you could probably do with a nema14 because you won't need a high torque because of the worm gear reduction. -Bas

yes indeed! Sorry for the late reply! It's awfully busy here... I think that the problem is that the printer can't handle the enormously high steps per unit that you would need because of the high gear reduction in the Flex3Drive worm gear set. I also noticed that you will run into trouble when using a UM2 extruder motor because it has 400 steps / rev and is on 16 microsteps. Erik van der Zalm also had this problem when he installed it on his UM2 so he adapted the firmware to do "double stepping". However because Erik has a very early version of the UM2 he has a 200 steps/rev extruder motor. Because of this it works just fine on his UM2 but not on other UM2s. The solution would be to install a 200 steps/rev extruder motor. I think we still have a bunch of them here at UM HQ. I'll check out and if we still have some lying around I'll send you one, that should solve your problem. cheers, -Bas

Hi everybody, At the moment I have Erik's design running on a UM2 However I did make a custom shaft for the hobbed wheel. The standard shaft that comes with the kit is about 10mm too short for Erik's design. At the moment we are experimenting with this. Kind regards, -Bas

Hi, I understand, that is of course unacceptable. I know that the current feeder design is critical, which means that only slight changes in the resistance in the "extrusion train" can make the difference between a nice and a bad print. From the material feeder perspective (which is the area of the printer I am working on) I can say that I am doing my best to get some more grip and torque. -Bas

Hi everybody! Just to give some input: I know that Illuminarti did some research about this as well: http://www.extrudable.me/2013/04/18/exploring-extrusion-variability-and-limits/ Besides this Erik van der Zalm also took some measurements with a rotary encoder on the filament (see attached graph). As far as I know, these results were all measured on a UM original. -Bas

Hi! I've never mounted a hobbed wheel onto a UM2, I've only tested it's grip in a test rig that I had built for that purpose. If your knurled wheel keeps getting loose you could secure it onto the motor shaft with a drop of loctite. Cheers, Bas

Hi everybody! Some input that might be interesting for those working on improving the feeder mechanism: Lately I have been experimenting a lot with geared stepper motors. The increased torque on the feeder axis helps a lot in achieving higher extrusion rates. However, printing with a very high number of retractions is still problematic: during these prints the filament gets flattened and obstructs the bowden tube. I have been measuring the temperature of the feeder wheel axis; during long prints (with a room temperature of 22 degC) the feeder wheel axis reaches temperatures between 40 and 43 degC. At this temperature PLA already gets quite a bit softer than at room temperature. During continuous extrusion this is not problematic, however during prints with lots of extrusions it is. Cheers! Bas

I did a quick comparison between the hobbed and the knurled feeder wheel. I fed a piece of PLA through a feeder while measuring the torque on the feeder motor axis with a torque sensor (no bowden tube and no hotend present). I used a hobbed wheel and a knurled wheel of the same diameter (8mm, for the hobbed wheel the diameter at the lowest part of the curved surface). At the same idler setting, a higher torque was needed to feed the material through with the hobbed wheel. However the hobbed wheel also had a higher grip than the knurled wheel at the same idler setting. So in the end it doesn't seem to make too much difference. -Bas

Hi! I've also noticed that the hobbed wheel seems to cut more into the filament while the knurled wheel seems to push and deform more. The only disadvantage that I found was that when the wheel really slips over the filament (e.g. in case of a clogged nozzle) then the grip pattern really clogs which dramatically decreases grip. However the grip was superior to the grip that I could reach with a knurled feeder wheel. I have the suspicion that the hobbed wheel is also a bit more efficient in transferring power to the filament because the grip pattern cuts better into the filament, resulting in lower losses. A while ago I've done some measurements on the grip versus the idler wheel pressure for different grip patterns, If you'd like I could send you the results. I've also done some measurements on the feeder loss versus the idler wheel pressure (the higher the pressure, the higher the grip, but also the higher the losses induced by deformation of the filament). I'll try to repeat this measurement this afternoon but then with a hobbed wheel. That way I can seen if the hobbed wheel is more efficient. Cheers! Bas

Hi Ian, One of us is experimenting with the design made by foehnsturm with the flex shaft. Cheers! Bas

Hi, I must say that I am really impressed by your design. Very cool that you've built this! I see that you use a hobbed extruder drive wheel and was wondering whether you've had any issues with the wheel clogging. I've done some tests with different types of grip patterns on the feeder wheel and the hobbed wheel performed best. However it also had the problem that it was prone to clogging. I was wondering whether you had any issues with clogging. Cheers, -Bas

Hi all, @ Ian: I've tried with a PU wheel, but that was way too slippery on the PLA. However, I did not try with other rubber wheels (PU wheels were most easy to buy). I must say that I am really impressed by the flexible shaft solution, seems to work pretty well! -Bas

I don't know about the levelling but I was wondering about something else. Did you do a factory reset after you changed the settings for the motor current? I've been experimenting with different motors for the feeder and if I remember correctly, the settings for motor current are stored in the EEPROM. When I changed the motor current setting in the firmware, the new settings were only implemented after I reset the printer to factory default.

Hi everybody, Thanks for the comments! I've added an updated version in which I made some minor changes based on the comments that I received (earlier through email). Thanks! Bas

Hi everybody, Several people asked me to publish research reports on the forum. Up until now I shared these reports through email. The advantage of this was that I had direct contact with the people who would read it. This meant that I received elaborate comments from almost all of them. Furthermore it was easy to keep track of who had read the report, so in case there would be a mistake in the report I could contact all of them personally. However, because people requested it, I will experiment with putting the reports online. Attached to this post is a report I wrote earlier about tests with different grip patterns on feeder wheels. Cheers! -Bas

Hi everybody! @Bertho & Robert: I've sent you a PM with a digital copy of the report. @Ian: I have tried to improve the existing feeder as much as possible within the boundaries of the current tooling. Besides that we are working on a new feeder. However before I started to design something I first wanted to determine the boundary conditions. Therefore I designed the test setup with the torque sensor and the setup in which I tested the different grip wheels. I'll send you a copy of the report with the torque sensor. I now have a rough idea about the amount of torque that I need, so now we can start working on the feeder redesign. This whole process of testing went quite a bit slower than I would have liked. This is because I am also quite busy with the other project that I have to work on. @Nicolinux: It might be that you still have the old feeder housing. I've attached a picture that shows how you can recognize the revised feeder housing. If you still have the old one, I'll make sure that you will get the revised feeder housing if you want.

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

Hi everybody, I saw that the github link for the step files of the revised UM2 feeder housing was not working anymore. For anyone who would like to print this revised version I have attached the stl files to this post. If you decide that you want to use this revised version, you'll need a metal grommet at the filament entrance. If you need one you can contact me. At the moment we are in the process of a toolchange for the feeder housing. I'll let you know when I have samples. Furthermore I have been running tests with different feeder wheels and I am currently in the process of designing a test bench in which we can do more detailed measurements on the resistance induced by the bowden tube and the hotend. Cheers! Bas

Hi Everybody! As Ian said I am indeed trying to solve the big obvious problems in a quick way (read: without having to build a completely new tool for injection molding). However, after that I would like to work on a new improved design of the feeder mechanism, so all ideas are welcome! In case anyone would like to take a closer look on the 3D models of the revised feeder housing, I’ve uploaded the STEP files to GitHub: https://github.com/Ultimaker/Ultimaker2/blob/master/1258/1258-B2P-D.STEP https://github.com/Ultimaker/Ultimaker2/blob/master/1259/1259-B2P-D.STEP : Metal grommet: I just ordered a bunch of those rings for testing purposes. Let me know how much you need of them and I’ll make sure that you’ll get them. Flexible filament: I really like the idea of a snap on part that can be used for printing with flexible filament. We discussed a similar idea with Ian. I will definitely keep it in mind as an upgrade. : Great that you would like to test the revised parts. Could you please send me a message or an email with your address? Then I’ll make sure that you will get a sample of the metal grommet. @Takei Naodar: Of course! You can download the STEP files from GitHub. : Thanks for sharing your ideas! As said I am at the moment trying to solve the obvious problems in a quick way, but I’ll definitely keep them in mind. : Embedding image: I also had some issues with pasting an image link in the post and got a message that it wasn’t allowed. In the end I included the image as an attachment to the post, which did work. Teflon Tube: The closer to the feeder wheel, the better I guess. Do you think that the tube will still have a well-defined “end point” in your setup? Thanks again for all your feedback! -Bas