As a shoemaker I am always looking for new fabrication techniques- there are hundreds of separate steps to make a shoe and many ways of doing each one. I have always wondered how else a sole might be attached to a shoe 'upper.'

Traditional forms of attaching shoe soles to material uppers have included glueing or stitching (sometimes both) of the upper material to the sole.

The glueing process is toxic and messy. The stitching process is heavyweight and very time consuming.

I really dislike using adhesives- neoprene is a common one for footwear. It contains some horrible stuff that can do some serious damage to people and wildlife. I have attached an outtake of a safety data sheet from one of the leading brands of shoemaking glue. Pretty depressing reading. During the process of attaching a pari of soles, you will inevitably come in to contact with it in some way, even with PPE. Add to this the fact that to make the glue adhere correctly, you need to abrade the leather of the shoe, thus causing particulate of leather to become ariborne and possibly inhaled. Having spent years working with the sticky stuff, I have begun exploring ideas about new ways to attach the sole to the upper without using adhesives.

The other benefit with having a modular shoe means that items can be interchanged, thus increasing longevity and allowing for easy customisation.

The design that I am outlining in this project is still in progress as I publish this Instructable. This is my method for creating a shoe without using glue, I encourage the reader to try and find their own ways also!

Step 1: Gathering Materials and Resources

For this experiment, whilst avoiding glue I will also try to use the fewest possible number of specific shoemaking tools or machines.

A basic list of things you will need:

Some material for the upper (leather/fabric/PU leather- the choice is yours)

A sewing machine

A 3D print printer (this assumes you have access to a computer and some software such as Fusion 360- free of charge to use if you use it for an enterprise earning less than 100k / year)

Some flexible filament - TPU

Cord or elastic

Cutting instruments

Step 2: Initial Explorations

My idea for a no-glue-shoe went through several design iterations before I settled on something that worked reasonably well.

My initial research and development focused on beginning to 3D print certain components of the shoe that would remove a step from the traditional shoemaking workflow

I started with making a heel cup. These are traditionally made by shaping leather pieces and using homemade water based glue to form them and place them between the lining of the shoe and the outer. More modern techniques include big machines to heat and mould synthetic materials into the desired shapes.

Coning from a traditional background I have been looking at ways to digitize my workflow. This means designing shoe lasts with computer aided modelling and 3D printing of components. I figured this project would be perfect to really test those skills.

I started by creating my shoe last in Autodesk Crispin Lastmaker software (https://www.instructables.com/id/Last-Comes-First/). This model can then be directly imported in to Fusion 360 software to model components from.

I started by modelling a relatively simple form of the heel cup and modelled holes though it so that I can attach this to the shoe upper using elastic cord.

The model was printed using Ninajtek SemiFlex filament on a Makerbot Replicator 2. As you can see the results are highly satisfactory for such an inexpensive machine. The filament seems perfect for shoe making applications.

Step 3: Integrated Heel Cup and Sole

Having successfully printed a heel cup, I decided to advance to trying a sole unit that has an integrated heel cup.

Using the same cord attachment for the sole and heel I decided to up the 3D modelling ante.

First, I used the same imported shoe last but this time also exported the sole piece from Crispin into Fusion 360. Having the sole bed as a 3D file allowed much greater ease for modelling- it meant I did not have to make the sole as a flat piece which would be wrapped to the last. This is important as I need it to hug the last shape as much as possible if im note using strong glue.

Adding the model of the heel cup and patterning eyelets along the sole edge I was able to create a system that would allow a simple weave to attache the shoe upper to it.

Again this model was made from SemiFlex material however this time printed on a homemade machine similar to a makerbot machine.

The result was again good and I see merit in the attachment system, however I do not like the appearance of it and think I need to increase the number of eyelets. Also there is a disparity of flexibility between the leather forepart and much stiffer heel part.

Step 4: Digital Welt

Traditional methods of shoemaking would use a stitched construction. This used to be done by hand and more recently by machine.

The leather upper is sewn once to a strip of leather, known as a 'welt'. This in turn is sewn to a leather sole. The result is a hard wearing and repairable shoe.

I figured I could employ some of these elements for my experiment.

By making my own welt I could allow the shoe upper to be secured to the sole via cord. I modelled a long strip with eyelets along it, then a flat sole piece with the same number of eyelets along it. The idea beign that holes wuld be cut in the leather to allow the eyelet throguh to match up with the sole

I printed these parts and ran some tests. unfortuantely altohugh the eyelet spacing was equal for both parts, when the shoe upper is lasted, certain areas are stretched around the curves of the shoe last. this measn that some eyelets would marry with their partenrt and some would be off, by as much as one eyelets length.

Also, the welt provides no support for the foot in terms of insole, so I was forced to glue (!!) one to the upper in order to conduct the experiment.

The design shows potential but needs re-working in order to get the eyelets to match up.

Step 5: Digital Welt V2

Looking at the previous two designs plus going back to look at images of the hand welting process, the idea struck me that I needed to include the welt to the insole, with a separate outsole. This more solid insole would stop any great mis-alignment of the eyelets and allow the pieces to marry together much more accurately.

I took the exact bottom profile from my Crispin last model and extruded this to create depth. I made the eyelets longer so that they would protrude clearly past the edge of the leather and allow for the cord to be woven easily through them.

By making an extrusion of this profile I also created the sole unit with greater depth to provide cushioning and support.

I printed this two part design in SemiFlex and then began to create a pattern for the shoe upper. As shoe leather will be stretched and pleated around a toe curve, I was not sure how to pattern this into a 2D hole punch guide. I found it impossible to come up with a 2D pattern for this from scratch. I was unable to think of a reasonable way to gauge where the holes needed to be in the leather without a very handmade method. So, I created a mock up shoe with no sole, then aligned the insole piece to my shoe upper. Using a silver pen I marked the spacing of the eyelets on the bottom of the shoe upper. Now I had the markings, I pulled the shoe pieces apart, flattened the leather and scanned the upper. I traced where I had marked the leather and drew holes in as vectors. I now had a pattern from which I could accurately laser cut holes for the eyelets to fit through. I also needed to make several versions of the patttern in order to dial-in the hole position and spacing.

Step 6: Developing the Shoe Upper

So my approach so far has been that of a traditional shoe maker who stumbled across a computer and a printer...

I made my basic shoe pattern using the traditional technique, see my Instructable bout pattern making here: https://www.instructables.com/id/Make-Your-Own-Shoe...

If you are lucky enough to be using Crispin Shoemaker, it will allow you to export a flattened DXF of the pattern you can work from.

My traditional pattern still had a big flap of leather, known as a lasting allowance, left underneath the shoe with nowhere to go. This also meant for early experiments i still used glue to trap these flaps underneath the insole, just so I could test and iterate quickly you understand.

I found in wearing the shoes the leather squeaked a great deal against the printed TPU- embarrassing for a shoemaker

I decided that if i trimmed just enough material off the bottom, the upper would stay attached to the sole and not tear.

As I am not using any lining material I have experminreted with varying thicknesses of leather to find a balance between strength and comfort.

I heard of a plug-in for Fusion 360 called ExactFlat. The theory being you can create any 3D model then flatten ot into 2D- perfect for footwear, upholstery etc.
I took my 3D last and drew style lines on to it. I removed these parts so that the model looke like the shoe I wanted to flatten. I then imported the sole with eyelets and married the two together to see where I require holes in the upper. This all went well until the final steps of the process where I realised that the computer software is not advance enough to take into account all the leather stretching, pleating and manipulation. However, the software is incredibly impressive and I am sure for a less exacting project it would have suited my needs just fine I got close. So close...

Step 7: Tread Designs and Assembly

As a shoemaker I always assumed that designing a sole pattern would be the most easy and fun job that a shoe person could take on. It turns out, it is actually quite tricky to come up with a new and original tread design from scratch...

In Fusion 360, I spent a lot of time trying out different designs- triangles, lines, thicker lines, more triangles etc. I was thrilled that I could prototype these on the Makerbot. I found that using NinjaFlex rather than SemiFlex made a grippier and more cushioned sole. I also found that 25% Hexagonal Infill worked very well for support, cushioning against material use and time.

For sole attachment I found that both 1/16th inch elastic cord worked as well as shoe laces. I did find that leather laces did not work as well as they wore through quickly and once stretched would not retain their original tension.

The assembly technique I have employed is a simple one that doesn't require a lot of cord and is reminiscent of the Greek 'meander' style design.

I used a solid shoe last for the assembly however it is possible to make the shoe without the last, you just get less definition of the toe shape.

NOTE: Using SemiFlex for the insole and NinjaFlex for the outsole allowed easier threading for the insole- sometimes it can be tricky to put the cord through the eyelet but when it is stiffer (SemiFlex) it can be slightly easier.

Step 8: Finished Design

I produced a range of different upper designs but my favourite is the 'ghilie' style shoe as the unique lacing design reminds me of the lace technique I have used to assemble the soles. The contrast colours give the shoe a fun and playful feel. The sole seems supportive, grippy and cushioning and the fastening mechanism appears to be quite robust. I will now go and test these on the street to see what improvements can be made and how durable the system is. I keeping an eye on how the TPU wears over time and how long it lasts before it needs to be swapped.

It feels great to have made a pair of shoes without using adhesives, but that look like 'real' shoes!

I would like to say a huge thanks to Andreas Bastian who's collaborative help provided a great deal of support and expertise in printing flexible filaments, amongst other things. You can check him out here; https://www.instructables.com/member/andreasbastian...