I'm setting up my printer in a small room, and I thought I'd come up with a system for more easily swapping filaments, but I'm not yet sure it is feasible.

Rather than physically replace the spool, I'd like to hang most of my spools on the wall, where they can rotate, and only swap filament leads in the (direct drive) extruder. The spools would be placed at some distance from the extruder, and at various angles. So to make sure the filament is pulled from the spool at the proper angle, and to avoid breaking it, I think it should probably run through a flexible tube to reach the printer.

I know such tubes are used for Bowden style extruders, but what I'm proposing is different in at least two ways: [1] the extruder motor would be pulling (not pushing) filament through the tube and [2] the tube would be longer than normal, e.g., between 1 and 2 meters.

Is this plan feasible? Or are the problems I am not foreseeing?

Edit 1: I don't know where I got 1-2 meters from. The distance is actually closer to 50cm.

Edit 2: To clarify my proposed setup, here's a crude mockup:


  • $\begingroup$ Fantastic approach and I really like the possibilities that this idea creates. My questions: How would the printer know to swap filaments, and which filaments to swap? (potentially written into the G-Code?) How will the printer swap the filament? In answer to the second question, I am envisioning some type of cutter which then melts (fuses) the second filament to the first, but how can this be accomplished in a uniform fashion? Anyways, I think this is a fantastic idea, I'll be watching this thread closely! Contact me if I can help in any way. $\endgroup$ – J. Roibal - BlockchainEng Jun 13 '16 at 21:21
  • $\begingroup$ @J.Roibal: Your interpretation is a bit more ambitious than my actual idea. :-) I was still thinking of swapping filaments manually, just minimizing the time and effort required. This would certainly help you to hot-swap filaments during a print, but I currently have no plans to fully automate this. --- I know about another recent project you may be interested in. But it's a bit more pricey than what I have in mind. $\endgroup$ – mhelvens Jun 13 '16 at 21:30
  • $\begingroup$ my heavens, mhelvens! I can totally see how this could become automated, removing the necessity of dual-extruder 3-d printers while adding the capability to print in dozens of colors and types of materials, automatically. $\endgroup$ – J. Roibal - BlockchainEng Jun 13 '16 at 21:32
  • $\begingroup$ @J.Roibal: Sure. I can see that too. And I may automate the process in the future. But it would be a hobby project, as this is, and I would not achieve the precision required for multi-color prints. Developing that technology would be a full-time effort, because it is more complicated than you might think. --- Anyway, as I hinted in my previous comment, the technology you want does already exist. Have a look at the Palette. I wish I could afford one. $\endgroup$ – mhelvens Jun 13 '16 at 22:16
  • $\begingroup$ @J.Roibal: The Palette splices multiple filaments into a single stream, as you suggest. The problem is that colors mix together in the hot-end. To get crisp color-swaps, they build a 'transition tower' on the side of the main print, to purge the mixed plastic. --- The solution we're talking about here would work differently. For every swap, it would retract the current filament, and start extruding another. This would probably evolve into multiple Bowden extruders feeding the same hot-end, and might not produce as much waste. Sounds like a fun project. But I do have a day job. ;-) $\endgroup$ – mhelvens Jun 13 '16 at 22:24

I would say it's not the best approach.

In the situation you've described I would rather install some kind of vertical ring a bit above your printer. This ring would organize your filament going from any direction.

So let's assume you got one spool on the right and one on the left of your printer, both are 1m away from the printer itself. If you get filament and put throught the ring then to your extruder.

In my opinion the ring should be around 15..20cm over the printer, it should have around 10..15cm diameter and spools should not be too low.

The only issue I see is unwinding spools which are too far from the printer because then filament would be unwinded angular but the same issue would be with and without teflon tube.

[edit] Here is a simple change of your design

enter image description here

[edit2] Please also note that you can assure proper angle of unwinding filament by installing small rings ("eyes") next to spools. These eyes will direct filament out of spool and then filament will turn to your big ring mounted on the printer. This is common solution in textile industry

enter image description here

| improve this answer | |
  • $\begingroup$ Thanx Tormod. Acording to my experience - tube longer than 1m is not a good idea. #1 it's harder to pull/push filament. #2 such long/heavy tube has influence on movement of the caret. #3 it's uncomfortable to install/uninstall filament $\endgroup$ – darth pixel Jun 13 '16 at 11:54
  • $\begingroup$ I absolutely agree, Darth! :-) $\endgroup$ – Tormod Haugene Jun 13 '16 at 12:12
  • $\begingroup$ I'm not sure I understand your proposed setup with the ring. Would it be too much trouble to ask for a quick mockup image (like the one I added to the question)? $\endgroup$ – mhelvens Jun 13 '16 at 18:22
  • $\begingroup$ And the issue of pulling filament at an angle (and expecting the spool to keep up) is the reason I'm suggesting tubes. If the filament angle is properly guided at the source, I don't see that as a problem. Thoughts? $\endgroup$ – mhelvens Jun 13 '16 at 18:30
  • $\begingroup$ @mhelvens please take a look on my post (edit). Please also note that rings can be used on both sides of the design (Thanx Tormod). You can apply little "eye" next to the spool which will assure the filament is unwinded at proper angle, then there could be a loooong way to the ring proposed by me :) (of course on that way there could be even more rings which will direct filament properly)... assuming you are pulling, which you do $\endgroup$ – darth pixel Jun 14 '16 at 4:55

This is somewhat similar to a question I asked a while back. Pay attention to Ryan Carlyle's response in the question I posted. His post essentially explains various ways that guide tubes may hinder extrusion due to drag. In regards to the length of your guide tubes, ensure that the tube is relatively straight.

If you are holding the filament above the machine, chances are that you wont really need the tubes anyways and the distance between the spool and the extruder won't matter either. The point of tube is to ensure the filament doesn't bind coming from the back of the machine, thus potentially clogging your extruder.

Update Based on the OP's drawing, I think it would benefit to change the design to drop the tubes all together. The tubes can provide many point of contact and managing so many lines at once may become problematic. So, possibly drop the tubes and intend on not drawing your filament unless they are fed into the extruder. You could install smaller ends of tubes at either end of assembly (one at the storage rack, the other above the extruder) to ensure the filament is being fed/pulled straight. The space between the storage rack and the machine doesn't matter. Here's a basic edit of your design that might be easier to manage. For kicks, here's a design I made for my own machine as a "quick-change" design.

enter image description here

| improve this answer | |
  • $\begingroup$ Even if the spools hang above the printer (which is the plan), I feel like things might go wrong if the extruder tries to pull filament from the spool at an awkward angle. Anyway, the answer you're referring to is useful. Thanks! $\endgroup$ – mhelvens Jun 13 '16 at 18:27
  • $\begingroup$ Thank you for uploading your sketch, it helped understand more of what you're trying to achieve! +1 $\endgroup$ – tbm0115 Jun 13 '16 at 19:44

I would like to add to the already great discussion, that this general approach is prone to filament degradation by water intake. That is, if you are using filament for which this is an issue (PLA as the most prominent example).

Also, your goal is to exchange more quickly, that is without (un-)mounting whole spools each time. Basically you want to have a multi-filament spool holder, if I understand correctly. If you design an enclosed box with a single outlet you could incorporate silica gel inside and a reduced effort to swap the filament. I would still avoid a nice free-hanging spool solution due to the degradation issue.

| improve this answer | |
  • $\begingroup$ A good point! An enclosed box would definitely be a good move. This also seems to be an argument in favor of tubes all the way, rather than short filament guides on both ends. --- I don't understand your "box with a single outlet" plan, though. I don't see the point, and it would probably complicate filament swapping, because I'd have to open up the spool enclosure to feed a new strand through the tube. Right now I'm still inclined to run a bunch of separate tubes to slightly above the extruder, like in my mockup image. $\endgroup$ – mhelvens Jun 14 '16 at 20:26
  • $\begingroup$ Yes, if you can get the tubes somewhat airtight or don't use the first part of the filament each time you swap, your idea works better. Why wouldn't you want the whole box very close to the printer so that you wouldn't need the individual tubes altogether? Having 5 tubes being moved by the carriage is additional load to the motor that doesn't benefit the print itself. $\endgroup$ – kamuro Jun 15 '16 at 12:54
  • $\begingroup$ The placement of the spools is purely a matter of constraints in the printer-room. It turns out the spools would have to be at a distance of about 50cm. --- Moreover, as J. Roibal has suggested in a comment, it may be fun to try to fully automate filament swapping some day. I'd definitely need individual tubes for that. Would multiple tubes really add significantly to the load? I'll have to experiment. $\endgroup$ – mhelvens Jun 15 '16 at 13:29
  • $\begingroup$ The load may depend on how the tubes are directed to the printer. I had a problem when building a repstrap, - as soon as the bowden tube was attached, the motor wouldn't move anymore. Truth be told, it was a tiny cd laser carriage drive motor, but but it moved flawlessly without the tube. You can easily find out how much you have to increase the current at the border to losing steps. I would guess that it doesn't influence your printer too much (watch ghosting), but if it can be easily avoided, why shouldn't one think about it when designing ;) Fully automated sounds like real fun ;) $\endgroup$ – kamuro Jun 15 '16 at 16:24

Yes, this is feasible. You should constrain the tube at both ends, this prevents the spool from pulling up on the extruder. This approach is known as reverse bowden setup.

| improve this answer | |
  • $\begingroup$ What do you mean by 'constrain the tube at both ends'? Are you simply suggesting that neither end should hang loose? Please have a quick look at the image I added to the question, to see if it matches your meaning. $\endgroup$ – mhelvens Jun 13 '16 at 18:23
  • $\begingroup$ Yes, this matches what I mean. Without the tube, the force required to get the spools to unwind is transferred to the extruder (possibly moving it around and decreasing print quality). With the tube in place, that force is instead applied to the mount of the tube. $\endgroup$ – Tom van der Zanden Jun 14 '16 at 14:05

This shouldn't be too much of an issue, as long as:

  1. You guide the filament properly to avoid breakage (at both ends as Tom mentions)
  2. You avoid too much friction between the pulling extruder motor and the spool itself.

You mention that you consider using a 1-2 meter teflon tube to guide the filament. Although this might work, long teflon tubes will increase the friction from pulling the filament. Furthermore, I really doubt having such a long tube will be beneficial at all; quite on the contrary, I think it could be rather impractical. Most likely, having simple guides at both ends (whether its tube stubs or rings) will suffice.

| improve this answer | |
  • $\begingroup$ Aha. Guides at both ends, rather than a tube all the way. I can see the sense in that. Now that I think about it, other answers (like darth pixel's) may also have hinted at this. $\endgroup$ – mhelvens Jun 13 '16 at 18:28
  • $\begingroup$ @mhelvens, I believe that both Tom and Darth hinted at this. Whether it's a tube or rings probably doesn't really matter that much. :-) $\endgroup$ – Tormod Haugene Jun 13 '16 at 19:51
  • 1
    $\begingroup$ @TormodHaugene Actually, tube v.s. rings makes a difference. A tube has the advantage I mentioned in my answer, rings don't. $\endgroup$ – Tom van der Zanden Jun 15 '16 at 10:32

Your idea is actually good, but i would add some edits to it. First of all, to keep your equipment clean, you would like to add little foam rubbers to clean the filament at the entrance of the tubes, to clean out the dust that will be accumulating over the rolls.

Another thing that i would recommend is to hang the tubes on a separate support for them and not on the extruder head, as it would add some additional friction to it and can cause the 3d model with additional friction making it distorted.

| improve this answer | |
  • $\begingroup$ Can you discuss your improvements further? I am interested in the maintenance portion of your answer, in particular. $\endgroup$ – J. Roibal - BlockchainEng Jun 13 '16 at 22:48
  • 1
    $\begingroup$ @DanBoyko: I'd be interested to see you finish that last sentence. :-) $\endgroup$ – mhelvens Jun 14 '16 at 20:22
  • $\begingroup$ i am really sorry, don't know why it cut the answer, i was about to say that i would damage the 3d model with additional friction making it distorted $\endgroup$ – Dan Boyko Jul 2 '16 at 0:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.