Sunday, 3 June 2012

Simplified multi-colour filament experiment - #30DoC day 3

I woke up today and had an idea for a simple way to join lots of tiny sections of plastic printing filament together, so I thought I would give it a try for Day 3 of  30 days of creativity ( #30DoC ).

Ever since my first filament joiner I have been thinking about easier ways to simplify the process. It's quite hard to join sections smaller than about 50mm manually and I always wanted a simple way to change with lots of colours every layer of an object, that would indeed look totally great!
Filament joiner here on Thingiverse if you fancy making one.

So my idea was to use PTFE tube (6mm outer and 3mm inner) to hold tiny cut sections of 3mm printing filament and pop the coiled tube in an oven to fuse the sections together, all sounds perfectly dandy and totally awesome? it should work? right?

Cutting sections of PLA and inserting into the tube was no problems at all.
I had a 1.2M coil of PTFE tube and lots of tiny sections of filament.

It coiled easily and fitted in the oven, heated up and the filament started to melt.

It has a little oozing at each end and I could see the sections were not fusing together as I had hoped.

Exactly the opposite was happening, they were moving apart from each other.

The PLA was getting nice and soft, unfortunately so was the PTFE tube. I removed it and tried to push a 2.5mm metal rod up the tube, this didn't work and I concluded the tube was just too long.

I cut the tube in half and popped it back in the oven, this time it worked a little and I had a section of joined filament out the other end. Then the PTFE tube cooled down and trapped my rod inside the tube stuck to the filament!

I had a few more attempts, but each time the sections that came out were not fused well enough together or the PTFE tube swelled and got jammed up.

I'm pretty sure that PTFE tube elongates as it heats up, and as my PLA sections were not constrained in the tube they also moved along the tube or generally decided to not want to fuse together and actually shrink in length while getting a little fatter in diameter.

So could blocking one end of the tube with a M3.5 bolt and then compressing the filament the other end as it heats up give a single filament? well worth a try at some point?

Or maybe the tube needs to be quickly cooled down so the contraction helps form the filament?

Or possibly the filament needs to be squashed down the tube while still hot and then allowed to cool. (I'm thinking that it will never come out again then)...

Or just maybe the hot tube and contents needs to be passed through a number of aligned sets of bearings squashing it and it's contents to allow fusion and also reducing the filament inside a little to give it a fighting chance to get the damn filament out of the tube again!

Thinking about it now as I type this, it's the same problem we get when the thermal insulator of a hot-end is above 60 degrees C, the PLA jambs up and refused to slide so it could just need to be mechanically stretched after fusing inside the tube while it's still all hot, the bearing die idea may actually work...

So it failed... but I still wanted to post the idea as someone out there may well be able to make it work.

It could have been great! If I get more time I'll try again, Ideas anyone?

Don't forget to take a look at all the other wonderful work for 30 days of creation, * be inspired *
#30DoC the story so far - Day 1 -    Day 2 -   Day 3 -

This is really going to bug me now...I was dreaming of printing a rainbow Giraffe for you all :(

You gotta fail a bit or you're not trying hard enough!

More tomorrow.




  1. I posted a similar idea, though less ambitious, on the makerbot list to use a piece of PTFE tube instead of your machined metal block to heat and shape the filament joint. I also suggested using ABS with acetone to do a solvent weld instead of heating it up.

    No matter how you do it, you'll definitely need some axial compression force to get the joint. I also suspect that even if you get good joints with your method, you'll have trouble getting the filament out due to the combination of nothing to grab at the ends and the very tight fit due to the filament forming itself to the inside of the tube while hot.

    One possible solution for the expansion of the PTFE is to put it inside a stainless steel tube. This is how MBI's Mk5/6 extruders work, and while you can't pull the filament out cold, ABS slides out pretty easily without deforming much at 170-180C (there's probably a different sweet spot for PLA, but I haven't tried it with that extruder).

    I'm thinking that you might be able to heat a short section of the tube while keeping the ends cold. You could then manually or mechanically feed a series of pieces down the tube where they would theoretically join in the hot section and maintain their shape as they cool while moving through the cold section. You would need to compress the hot section so that its diameter would be slightly less than the cold section instead of slightly more due to thermal expansion. Keep it moving and it shouldn't sieze up. This method should also be useful for getting a consistent diameter when making new feedstock out of nurdles, scraps, and/or shredded pieces of old prints to be recycled.

    1. Thanks, I like the ideas, maybe instead of Stainless use sections of clamped aluminium block around the PTFE tube, wrap nichrome around them or use resistors and have the tube sections go from very hot to melt the filament sections down to a bit cooler to allow flow and movement.
      As you say keep the ends cold and try to push/pull out the new joined filament.

      It would be neat if this sort of thing could also form the basis for a Scrap - to - feedstock recycler/former, I wonder if pulling on the cool formed filament end would produce enough suction to pull down molten PLA into the tube behind it, may need a pressurised feed chamber to help it along.

      I did find a Teflon funnel the other day...

      It's going to need some more investigation, but sounds interesting and possible.



    2. I played with the idea of a temp controlled heated razor blade (and a heated thin plate) that cuts and/or heats both sides of the filament. You really only want only the contact area heated and welded, not the whole filament. If the heat affected zone is too cold, you get a cold weld that is very brittle. If it is too hot, it deforms the filament by changing either the diameter or the angle of one filament segment to the other.

      After it is welded (but still very soft), I have been trying to shove it and pull it through a metal "sizing tube / hole" that fixes any oversized diameters that will get jammed in the extruder. I might try your PTFE filament tube to act as a pre sizer that massages the filament smaller before it hits the metal sizing tube. I'm only trying to get a very small amount of plastic melted, so I can keep the rest of the filament above the glass transision temp where it is very rigid. The smaller the heat affected zone, the quicker it cools. Plastic is a very poor conductor of heat.

      Take Care,

  2. I've just read this now, and I've come up with a solution to your problem. If you use your original filament joiner to join filament before cutting it then you can have the extra length needed to easily join it. After that you can cut the filament to leave the right length of plastic joined. This could give you the fast colour changes you want. Once my printer is working I'll try it out.