3D Printed Stretch Fabric Seams

This instructable documents an experiment to combine stretch knit fabrics (in this case powermesh) with 3d printed mesh to make stretch seams using an extruder printer. I want to find a way to insert 3D printed panels into stretch wear for etextiles.

I am a total 3D print novice and my background is in garment construction. I wanted to learn both how to use the printer and at the same time to apply it to the type of things I usually create. I think this approach to 3D printing is a bit unconventional but that is the benefit of being a novice. You make it up as you go along!

It is a hack in which I start then pause the printer and lay the fabric in place, then continue the print. After a few earlier experiments trying unsuccessfully to slice in the print file I decided this start/pause method was the simplest way to create a first sample to see the materials together and make a proof of concept.

It is very hands on and time based idea. You have to watch the printer and be quick with the steps and the failure rate is initially high while you test. Sometimes it is better to hack a process and test the materials to discover the best way to proceed before trying a longer and more time invested approach. I wanted to try this quickly and experiment with pushing the printer capabilities to the max!

The eventual goal is to make garments which have three dimensional elements printed into the garment itself.

Fabric: Powermesh because it is a super stretchy four way stretch material. It is a great construction fabric for garments because it is very strong. It is also super flat and smooth, which works well on the 3D printer bed.

Filament: Ninjaflex Semi flex filament in black and clear.
I decided to use Semiflex rather than full on Ninjaflex as it is a little more stable to use and easier to print with. The black seems more soft and the clear seems a little more brittle so I thought it was a good idea to try a sample for each, for both aesthetics (which is important for my project) as well as their physical properties.

Printer: I used a deconstructed Makerbot Rep2 and no special settings, besides the 220 temperature settings for semi-flex filament. This is because the Makerbot is a direct feed machine (it uses no long tubes like the Ultimaker) and so you have more control as the filament feeds into the machine. I used the machine with all the side walls removed so I could get my hands in at three sides of the bed during the printing process. This is a hands on hack!

First thing is to find or make the 3D file to use. I am not a 3D modeller so I was lucky enough to be given a file I could adapt. I had already seen a test print so I knew it was basically the right starting point.

I took an existing file by Andreas Bastian of a tried and tested printed mesh. Because of the construction of the model, the print already has the ability to stretch in one direction. The semi-flex filament also allows a little stretch. This means it is stabilised in one direction and is less likely to tear once printed to the fabric.

The printed mesh is bigger in size than the fabric mesh I am using but it's good as a starting point. If i had matched the print scale to the fabric the print wouldn't stretch much at all. The larger holes in the print are probably a good balance for the stretch of the fabric. This is just a first test, after all. I can tweak later.

The file I have uploaded is the source file I adapted. Instructables wouldn't let me upload a .x3g makerbot or a .thing file. Sorry!

Prep

Mark up the print bed:
First do a test print to see the file and also figure out where the print will start. I printed the first few layers of the file and cancelled the print. Four or five layers seemed a sturdy thickness and compared well to the fabric.

Prepare the fabric: I cut the fabric larger than the print size so there is space outside of the print area to secure the fabric with tape.

I placed my fabric overlapping the outside edge of the print by 1.5cm and marked points outside the print area with tape so i could accurately place the fabric during the print.

Printing

For my first test I printed half of the mesh file in semiflex and cancelled the print when it looked thick enough.

I printed three layers of the mesh print before pausing the print (when the extruder was at the far side of the bed, away from the seam) and quickly placed the fabric and tape.

I laid the fabric so it overlapped the edge of the printed area by 1.5cm. I stuck it down with masking tape just outside of the print area on three sides so the material was tensioned and stabilised as much as possible but so the filament would not print on the tape.

I needed to place the fabric quickly and continue the print so the filament didn't cool down in the printer. I wanted be able to continue the print without making too much of a mess! I then restarted the print and crossed my fingers that the extruder wouldn't snag.

It turned out that even three layers before placing the fabric was far too much. It meant the fabric was high on the bed and made the extruder mess up the first part of the print making the seam look messy. It did work though, despite not being very neat.

The finished result was stable but a bit too think. The print overwhelmed the fabric.

After a few tests I had figured out a method.

When printing you need to make sure the extruder starts on the fabric (left in the images) and moves towards the edge. This means the fabric is secured to the print in a place where it is more stable (close to the tape) and so it doesn't snag as easily.

I found it most effective this way, otherwise the fabric would be pushed and folded by the extruder. In the images above you can see the beginning of the extruder path sits 1.5cm in on the mesh and is printing towards the edge of the fabric.

To make the seam work you need to balance the amount of layers in relation to the thickness of the fabric. The fabric also needs to be really thin! If the print is too thick it does not feel like a viable fabric and does not move well with the material. If the print is too thin is falls apart.

Here are the stages i used:

- prepare and test to print from left to right (see images)
- print two layers of the mesh
- pause the print before it completes the layer so the extruder is far away from your seam
- quickly lay the fabric in place, overlapping the edge of the print by about 1/2 inch or 1.5cm and placing tape close to the edges of the seam line
- make sure the fabric is as flat as possible, especially at the point where the extruder will travel back across the bed to start a new print layer. This is the main point where the print will fail. If you're smart you could program the extruder to travel back across the tape. I need to learn how to do this!
- restart the print, making sure the extruder starts on the fabric and progresses off the edge, away from the fabric rather than towards it
- print two or three more layers

I used a thicker mesh file than required and just stopped the print when I felt the fabric was thick enough, rather than trying to change the mesh file each time. This meant I could change my print thickness depending on the fabric.

I made quite a few samples in both the black and clear semiflex. The black definitely worked the best as it was less brittle than the clear stuff which didn't feel as nice and had a tendency to just crack if it wasn't printed thick enough. The clear needed an extra layer to be strong enough meaning it was always thicker on one side of the seam.

These tests were just a small straight seam and in each instance I made them larger then required and trimmed the final samples because the edges were often a little messy. This is not ideal but when you make a garment there are usually seam allowances, so this felt similar.

I wanted to learn how the materials felt together so I could then go on to devise a better way to slice the print file to make it work more efficiently and reliably. This method relies hugely on the dexterity of the human and the ability to place fabric accurately at speed. One slip up and the sample is ruined. It's a collaboration between user and machine but the aim is to make the process more reliably mechanised.

When I started I experimented with various kinds of net and powermesh. This sample was a thinner and softer material and just didn't quite have enough structure so the nozzle on the extruder pushed it and warped the surface too much.

It made sense because really I wanted to use a fabric with a similar structure to the print to create a balance, the same as you would do in any garment. So i decided to go with the stronger material for all my subsequent prints.

The images above are also what happens if the extruder snags the fabric initially and disturbs the flat surface: the sample is a goner. There were a few failed samples like this, but generally I got the idea working efficiently enough to mass produce 30 perfect samples with only one error in the batch! These were for a swatchbook exchange where a bunch of us share tactile samples of physical experiments

I concluded this is a pretty good starting point to hack 3D printing for fabrics. It's got a high initial failure rate and is very hands on so it would need a lot more playing about to find a stable way to develop it as a method. However it seems lots of 3D printing has quite a high failure rate, so I guess that's not too unexpected. It is a quick and fun way to see how knit fabrics and printed fabrics interact.

I still have questions such as:
Q. can i skip a layer where the fabric sits so i can use thicker fabric?
A. Not without getting some custom code. This will require assistance from a smart person who codes! Though I did not need this on the Makerbot. As long as the fabric was perfectly flat and didn't initially snag, the extruder pushed the fabric into the material and this made a good bond.

I discovered that slicing files and leaving gaps or starting above the bed is not something which you can get a standard printer program to easily do. I suppose because it's not really called for. It seems you have to trick the print file to do something like print little towers at the edge of the bed if you want it to skip layers and leave space to inlay another material. So in review, my hack method was actually quite a sensible approach to get the test done quick.

Next I decided I would experiment with bonding other stretch material to my print. I have been experimenting a lot with bonded conductive fabrics using stretch bonding materials. For more info on this method (shown in the photos0 take a look at my stretch circuit instructable.

I think my fabric choice (powermesh) was successful because it was synthetic material. This meant it melted into the print.

I discovered that gently heat bonding these conductive traces to the mesh works beautifully. I used a pressing cloth to protect the materials and the gentle heat of the iron made the traces sink into the filament making the traces flush with the rest of the material.

I also tried weaving the conductive traces through the mesh, cutting my traces to 4mm to match the holes in the mesh. this would be a great way to add traces to a garment after it was made, or to be able to alter a circuit in an etextile garment once it is assembled. the woven traces could also be heated to secure them permanently in place. This is a sort of hybrid approach, taking advantage of the low melting point of the filament alongside synthetic fabrics. think this is a good way to go as its fun and sculptural and means i can use small pieces of failed test print and melt them together with the iron!

The eventual goal is to make garments which have three dimensional printed elements printed into the garment itself. So next I will try to print piece which is fully surrounded by fabric, the basis for a pouch shown in the sketch above.