<![CDATA[Gyrobot Ltd - CAD Design and 3D Printing

<![CDATA[Gyrobot Ltd - CAD Design and 3D Printing - Blog]]>Sun, 26 Apr 2026 16:42:15 +0100Weebly<![CDATA[Gensole is Live - Hello World!]]>Sat, 11 Jun 2016 13:29:15 GMThttp://www.gyrobot.co.uk/blog/gensole-is-live-hello-worldIt has been a long time since my last post on 3D printed insoles or indeed any other subject, but there is good reason.

Over the past year I have taken all I know (insert joke here) on designing and 3D printing insoles and packaged that up into a mostly all-in-one piece of software called Gensole®.

What is Gensole?

Gensole is a browser based tool which allows you to design insoles, optimised for 3D printing using the FDM process and Filaflex (TPE/TPU) type materials. It will ultimately export .AMF files for importing into Slic3r for 3D printer toolpath creation.

It is free for makers, and researchers for non-commercial use :

Who is Gensole?

Gensole is brought to you by Gyrobot (AKA, Steve Wood), an expert in flexible FDM 3D printing and the UK beta tester of Filaflex 3D printing filament and related products from Recreus. Credit to Damian Axford and Robert Longbottom from our Swindon Makerspace for their help with development. A mention to Footworx Podiatry Clinics who have offered help and support during software testing.

Why Gensole?

There has been a lot of development work with 3D printing insoles in the past couple of years by Gyrobot, and now is the time to separate a branch of that work out to a dedicated site. Gensole is the resulting package that takes all that knowledge and wraps it up into an easy to use software design and print solution. If you would like to read some of the background blogs and press write-ups, then please check out some of my previous posts or the links on the press page.

Gensole is the merger of four different insole processes that Gyrobot has looked into :

Form Fitting Insole.

Shaping of the upper surface of the insole to match a scan of the foot, using a brand new process called "Solemorph" :

Variable Density Insoles.

Areas of the insole printed with different flexible mesh densities to help offload high pressures:

Open/Closed Core Insoles.

Upper and lower surfaces of insole can optionally be printed with perforated holes to help airflow and stimulation of blood cells:

Shoe Profiling.

Parametrised insole boundary curves which are adjustable to provide a best fit inside shoes:

There are also a couple of YouTube videos to compliment the Instructions :

There is no ability to add orthotic corrections to the insole, however Gensole will match to a footscan, therefore change the footscan first and you can introduce corrective properties. The following video is providing ideas for the professional podiatrist who will know what is required when deforming the footscan :

All that said, now head over to gensole.com

]]><![CDATA[My Adventures with 3D Printed Insoles - Part 4/4]]>Fri, 02 Jan 2015 17:57:41 GMThttp://www.gyrobot.co.uk/blog/my-adventures-with-3d-printed-insoles-part-4-4Part 4 - Form Fitting Insole

...Continued from Part 3 - Variable Density Insole

Now it gets interesting. This post concentrates on the scanned data and image manipulation using Gimp, Inkscape and OpenSCAD, all of which are free and open source. However if further refinement of the 3D formed surface is required then Meshmixer is a great free tool (not opensource) for smoothing or twisting and pushing/pulling the mesh around.

I will show how to scan your foot and turn this into the CAD data used in the previous variable density insole to define the density zones, and also to create the surface data to fit the form of the sole of your foot. The form fitting surface is actually the easy part due to the new great import option in OpenSCAD which build a height map according to the brightness values in an imported digital photo. With a foot scanned on a simple document scanner the areas of the foot that are close to the scanning head, the brighter the image. The areas that are further away are darker. When the surface is built, the darker areas are thicker and the lighter areas are thinner, making a surface as shown :

Form Fitting Insole

Form Fitting Process

Step 1 - Scan the Foot

Capture foot with a flat-bed document scanner (A4 or A3).
IMPORTANT 1 : For image quality and safety, TRY NOT TO PUT PRESSURE ON SCANNER.
IMPORTANT 2 : Scan the foot in a dark room at night so the background of the scanned image is black.

Warning - HOBBIT foot alert

Step 2 - Manipulate the Image

Depending on the quality of the scanner and image it maybe necessary in Gimp to modify the Brightness (-75) and Contrast (+75). The idea is to try and achieve a dark background with the foot fading away into the distance, so experiment with the values somewhat.

Flip the image vertically, otherwise the finished insole will end up for the opposite foot.

Blur the image about 5 times in Gimp to make it have a "soft focus" effect.

Check the scale of the scanned image in Gimp and adjust if necessary. If you have an A4 scanned image make sure the Print Size is set to mm and the Width and Height are set to 297mm x 210mm resp.

It is most probable that the scanned resolution of the image is far too high for OpenSCAD to render quickly. One pixel of image will equal 1 unit of measurement in OpenSCAD (usually 1mm). If we have a high resolution scan then the resultant model will be both large in spacial terms as well as file storage and memory usage. Therefore we need to export out of GIMP a smaller less detailed image of the scan just for the surface form.

Scale the image so that the Width and Height = 297px x 210px and the X Y resolution is set to 1 px/mm each.

File then Export as a Scanned_Foot_Filename.png file, where "Scanned_Foot_Filename" is your choice, but we will refer to this filename later.

Step 3 - Creating the Surface in OpenSCAD

At the time of writing, the released version of OpenSCAD (2014.03) does not contain the latest command that is required for the surface creation and also the Offset command that we will touch on later. I downloaded a development snapshot of OpenSCAD, Version 2014.12.04. I guess any of the later versions will equally work.

This simple OpenSCAD code will create a surface solid, replace "Scanned_Foot_Filename" with the name of the scaled down file you created in the previous step :

scale([1, 1, 0.15])
surface(file = "Scanned_Foot_Filename.png", center = false, invert = true);

Z Scale at 0.05

Z Scale at 0.15

We need to crop this block to the profile of the foot from the scan. It is now where we can manipulate the image further. This will enable us to define the zones which we will also use in the previous Variable Density Insole post and also to create the profiling solid (often called a "cookie cutter") for this surface block.

I will therefore take a backwards move and re-visit a little more image manipulation in GIMP and the extra work in Inkscape for the Variable Density Meshes.

Variable Density Process

Step 1 - Artistic Cutout in GIMP

To achieve nice smooth zones, we start with the higher resolution blurred image in GIMP from the previous steps.

The great part of opensource is that many talented people all contribute to making the software better in one form or another. In the case of GIMP there are many plugins/extensions/scripts that are user created that offer extra effects that are not packaged within the original distribution of the released GIMP. There is an "Artistic Cutout" script that modifies the original image to a reduced number of colours and zones them off.

The Artistic Cutout script (originally by Eddy Verlinden) can be downloaded from here. It's a case of saving the script into the GIMP scripts installation folder so it can be used, more info here.

Once the script is available in GIMP choose the amount of zones (colours) you wish to print, try about 5 or 6, Note, the background black will be one of the zones, but like in this instance it could form one of the internal zones too (around the toes).

Open up the Toolbox (Ctrl+B) and Flood fill the background to turn it white.

This is the stage to make any minor tweaks or adjustments. Use the dropper to select a source zone colour of your choice then paint out any areas not required or smooth some boundaries, or just leave as is.

Export the image as a .bmp for importing into Inkscape.

Step 2 - Creating the Vector profiles in Inkscape

The image so far looks like it is suitably zoned but the data needs to be turned into profile curves and ultimately STL 3D data for slicing in our 3D printing programme. For this we import the previous exported .bmp file into the vector graphics editor Inkscape. At the time of writing I used Inkscape 0.48.5.

Open up Inkscape and set the document size to A4 Landscape and mm units, same size as the scanned image:

Import the .bmp image and set the size and position of the image to the document boundaries.

A great feature of Inkscape is that is will trace an imported bmp image and create vectorised zone patches which can then be used to export profile curves for use in a CAD programme such as OpenSCAD.

Set the Dialog box to the following options. You may need to change the number of scans until the preview (click update each time) matches the imported .bmp. Sometimes some zones will disappear if set too low and if set too high, extra zones that are not needed will be created.

Once the zones are created, they are automatically grouped together, so we need to Ungroup them to separate the paths.

If we zoom in (+ key) there always seems to be an erroneous zone created which we do not need so while we are zoomed in, we can select this and press the delete key to remove.

Before

After

Do not move the objects, but for example if we were to drag the objects about, you will see that separate zones have been created:

Example exploded view only, do not move the objects.

I would recommend that you now save your Inkscape work before we start to export each different object to a separate .dxf file. For this we need another plugin which allows the creation of a DXF file that OpenSCAD can read in (i.e. no splines).

The "better dxf output" plugin by Bob Cook can be downloaded from here. Usually extract the contents of the zip file into the Inkscape extensions installation folder (C:\Program Files\Inkscape\share\extensions\ for Windows). More information on the installation here.

Start by deleting all the objects except for the with the darkest zone, then save a copy of this as Left-1.dxf using the "Better DXF Output" file type option. Undo your deletes (Control+z) and then delete all zones except for the next darkest. Save this as Left-2.dxf, and do the same for all other zones like the following :

OpenSCAD Modelling

Form Fitting

Now that all of the profile curve files have been created for each zone, we can import these into OpenSCAD to create 3D stl files for use in our printer slicing programme (Slic3r). We also have a boundary curve (Left-5.dxf) which can be used to profile off the surface form insole that we created earlier in this post.

BUG ALERT!
I don't know why this happens but the dxf files that get exported from Inkscape seem to have their co-ordinates shifted by a random amount for each scanned image. It is this reason why the imported profiles and the surface data do not line up. I have therefore include an xoffset and a yoffset parameter which can be modified to align the profile solid over the surface, a bit of trial and error this one by pressing F5 to preview between each iteration of the offset values.

The two "#" at the beginning of the code below turn on the solid visualisation (translucent pink) for that command. When you see the two solids visually intersect then these can be removed to see the complete render.

Form Fitting OpenSCAD Code
User Variables in Bold

//
// This will turn a scanned foot image into a zheight surface and then profile clipped.
//
dxf_profile = "Left-5.dxf";
dxf_profile_offset = 3;
scanned_image_file = "Scanned_Foot_Filename.png";
zscale = 0.15; // Varies the height exaggeration
zinvert = true; // false will create a male version of the imported image
xoffset=-2; // BUG : Don't know why the DXF files have been offset by this random amount
yoffset=87; // BUG : Don't know why the DXF files have been offset by this random amount

intersection ()
{
translate ([xoffset,yoffset,-20])
{
# linear_extrude (100)
{
offset(dxf_profile_offset, join_type = "round")
{
import(file = dxf_profile, convexity = 12);
}
}
}
scale([1,1, zscale])
# surface(file = scanned_image_file, center = false, invert = zinvert);
}

When you are happy with the alignment and it looks something like the image below, press F6 to perform the Render (slow) and export an STL file. Also remember the X and Y offsets, we will use them in the Variable Density Insole below too.

Surface Form insole trimmed with profile curve solid.

Variable Density

The X and Y offsets from above will be used to position the dxf files in this OpenSCAD code too.

Use F5 for a quick preview with colours rendered. This option will vary the height of the solids so that you can see the visible zones. For exporting, the heights of the solid get modified so that they all sit inside the main shell, which is the purpose of the "stl_zoffset" parameter. When you are ready to export the solids, replace the export variable from "none" with the name of the profile you would like to export (e.g. "dxf_file1", "dxf_file2" etc.) then press F6 and export an STL file. Do this for each solid and you are now ready to program these variable density meshes.

Please see my previous blog post on how to programme these STL files.

Variable Density Solids created in OpenSCAD

Variable Density OpenSCAD Code
User Variables in Bold

//
// This will create the stls for the Variable Density zones.
//
//
// DXF files in order of Density : Hard = 1st in list.
//
dxf_file1 = "Left-1.dxf";
dxf_file2 = "Left-2.dxf";
dxf_file3 = "Left-3.dxf";
dxf_file4 = "Left-4.dxf";
dxf_file5 = "Left-5.dxf";
dxf_profile = "Left-5.dxf";
export = "none" ; // Set to "none" for F5 Colour visualisation, or "dxf_file[1-5]" above for exporting
export_height = 10 ; // Thickness of Insole Required
//
xoffset=-2; // BUG : Don't know why the DXF files have been offset by this random amount
yoffset=87; // BUG : Don't know why the DXF files have been offset by this random amount
stl_zoffset = 0.7 ; // Z Offset for internal stl profiles to achieve solid top/btm surfaces
profile_thickness = 0.9 ; // Sets thickness of perimeter "dxf_profile"
//
// Main Code
//
ztranslate = export=="none" ? 0 : stl_zoffset;
stl_zextrude = export=="none" ? 0.25 : 0;

translate([xoffset,yoffset,])
{
linear_extrude (export_height-(2*stl_zextrude)-(3*stl_zextrude))
{
offset(delta = profile_thickness, join_type = "round")
{
if (export == "dxf_profile" || export == "none") import(file = dxf_profile, convexity = 12);
}
}
}
translate([xoffset,yoffset,ztranslate])
{
color([0.04705,0.01960,0.01176])
linear_extrude (export_height-(2*stl_zoffset)+(5*stl_zextrude))
{
if (export == "dxf_file1" || export == "none") import(file = dxf_file1, convexity = 12);
}
color([0.2156862745098039,0.1411764705882353,0.1490196078431373])
linear_extrude (export_height-(2*stl_zoffset)+(4*stl_zextrude))
{
if (export == "dxf_file2" || export == "none") import(file = dxf_file2, convexity = 12);
}
color([0.4313725490196078,0.2705882352941176,0.2862745098039216])
linear_extrude (export_height-(2*stl_zoffset)+(3*stl_zextrude))
{
if (export == "dxf_file3" || export == "none") import(file = dxf_file3, convexity = 12);
}
color([0.7529411764705882,0.4745098039215686,0.4901960784313725])
linear_extrude (export_height-(2*stl_zoffset)+(2*stl_zextrude))
{
if (export == "dxf_file4" || export == "none") import(file = dxf_file4, convexity = 12);
}
color([0.9372549019607843,0.6509803921568627,0.6784313725490196])
linear_extrude (export_height-(2*stl_zoffset)+(1*stl_zextrude))
{
if (export == "dxf_file5" || export == "none") import(file = dxf_file5, convexity = 12);
}
}

Final Word

I realise this is a multi stage process and there is a lot of detail within this post, there maybe a way to reduce some of the steps or even somebody out there is dedicated and talented enough to automate some of the steps. I would love to hear improvements, issues advice etc.

Smoothing

One improvement I would like to suggest is to neutralise the natural colour variation of the underside of the foot. For this maybe a neutral coloured sock is preferred or an even coat of none sticky foundation make-up etc. This should help smooth out the Form Fitting surface somewhat and reduce the bumpiness. However I will now show you in Meshmixer (free tool) how it is a simple process to smooth out the insole surface.

Import the STL mesh into Meshmixer and set the render options with the spacebar (green circles):

Under the Edit menu, there is a "Generate Face Groups" option. This will help us easily select the upper surface for smoothing without selecting the side walls. Set the "Angle Thresh" to around 50 or similar so the upper surface gets a colour all of its own and then Accept.

Click the select icon on the left and double click the upper surface to select all of the triangles as a single face group. Once the upper surface turns orange, you can now choose to increase the insole thickness by going into the Deform - Transform menu or straight into the the Deform - Smooth (Control+F) menu (next step). If you choose to increase the thickness, just drag the vertical Z axis handle upwards until desired thickness is achieved.

Setting the Scale to around 20 smooths out the ripples. Accept the value that you are happy with and then Export out a new STL file.

You can download this model from Thingiverse, but it is probably useless for you, however now you can go and 3D Print your own insole/sandal/flipflop in Filaflex too.

]]><![CDATA[My Adventures with 3D Printed Insoles - Part 3/4]]>Mon, 22 Dec 2014 12:02:12 GMThttp://www.gyrobot.co.uk/blog/my-adventures-with-3d-printed-insoles-part-3-4Part 3 - Variable Density Insole

...Continued from Part 2 - Open Core Mesh

It is now where we start to see a major advantage of 3D printing over existing insole technology. The traditional method to make a custom formed insole is to CNC machine the insole out of a solid block of a suitable chosen hardness of material. Post machining modifications can be implemented by added inserts in different materials etc. Any operation that requires multiple steps is adding to the complexity, inventory of different material stocks, skills and of course cost.

Here we can see that for a single use of material and machine (3D printer), a complex structure of insole can be created with differing grades of hardness according to the recipients needs.

Insole consisting of firm and soft areas

This model was created using a scan of a recipients foot, the whole scanning and modelling procedure will be detailed in the next blog post.

The example above is a completely uniform 10mm thick for no other reason than to show you when holding it up to the light, you can see the variation of the density without any thickness changes distorting the light pattern.

I printed out some sample parts to try out the various infill percentages, both with top/btm surfaces and open core variants described in my previous post. These are great to squeeze in the hand to get a feel for the softness and flexibility of the material.

Uniform 10mm thick

Flexible and adaptable stiffness

Various infill percentages

One of the newer features in Slic3r is "Modifier Meshes", a great example where and how they can be used can be found on the original Slic3r blog. Please read this first to let yourself in gently.

My example stl files and Slic3r setting files can be downloaded from Thingiverse.

Note : Slic3r version used 1.1.7

If you import my Config Bundle and .amf file (Additive Manufacturing File) into Slic3r then the following settings are defined for you, allowing you to export gcode straight away.

How to Programme Variable Densities in Slic3r

1 - Add 1.stl

2 - Set Perimeters, Top and Btm Layers to 0

3 - Set Fill Density to same value as the next instruction

4 - Settings Tab, Load Modifier 1.stl and add the modifiers shown

5 - Settings Tab, Load Modifier 2.stl and add "Fill Density" modifier

6 - Settings Tab, Load Modifier 3.stl and add "Fill Density" modifier

7 - Settings Tab, Load Modifier 4.stl and add "Fill Density" modifier

8 - Settings Tab, Load Modifier PROFILE.stl and add "Perimeters" modifier

Export G-code as normal from the button on panel 1. As a backup, it is best to export your config bundle and .amf files so you can save your slicing settings for another session.

If all goes well, you should have some G-code that you can print with, you can check mine out here.

Variable Density G-code

Halfway through the print

It's important not to have any perimeters around these "Density Zones" otherwise this will create contours which will be firmer than the surrounding infill and could lead to discomfort. For this reason, perimeters were switched off in stage 2, however we still need some perimeters around the whole insole, so this is the reason for adding a PROFILE.stl in stage 8.

If you would like to have a practice at printing some variable density insoles, and messing around with the values then download the examples from my Thingiverse page.

So far, I have purposely stayed away from the techniques and procedures involved in the image and CAD side for simplicity until the next article. The free opensource tool chain used will be Gimp, Inkscape and finally OpenSCAD to create the CAD models. The result will offer "Form Fitting" variant too.

The disclaimer here is that I am only providing information, this is not medically certified and is not designed to be such, but hey in many parts of the world, people walk round on planks of wood or old car tyres so everything is builder beware. Go forth and MAKE!!!

Read Part 4 - Form Fitting Insole here.

]]><![CDATA[My Adventures with 3D Printed Insoles - Part 2/4]]>Wed, 17 Dec 2014 13:07:24 GMThttp://www.gyrobot.co.uk/blog/my-adventures-with-3d-printed-insoles-part-2-4Part 2 - Open Core Mesh

...Continued from Part 1 - Introduction

When I printed my first insole I noticed it was extremely tough, rugged, and elastic, these were all great properties for long term durability. However I noticed that the insole was relatively watertight and so I thought this would be an issue for perspiration and therefore personal hygiene.

Closed Core (Top and Bottom Surfaces Printed)

Open Core (Top and Bottom Surfaces Omitted)

As can be seen in the photograph above, the closed top and bottom surfaces effectively seals the insole. This option of insole although has been favourably received by Podiatrists, it would more likely need a cover adhered to the top surface to provide appropriate moisture absorption and some ventilation. This cover over time could need replacing/cleaning and add to the maintenance of this print.

This is the same model although simply mirrored for the opposite foot in this case. The top and bottom surfaces of the print have been omitted in the slicing software (Slic3r or CraftWare). This exposes the infill inside and allows some breath-ability to the sole of the foot. This insole can be fully washed in soapy water and sprayed with an antibacterial for longevity between washes.

For the same infill settings as the closed core, the insole is softer and is much quicker to print and uses less filament. A firmer print can be achieved by increasing the density of the infill.

Types of Infill

The infill pattern/density and orientation all change the properties of the insole, and this is where the skill of knowing what the recipients requirements are would determine what slicing parameters to select. There is of course the good old fashioned trial and error method.

Triangular Infill Pattern.

A triangular or honeycomb infill pattern such as the two photos above will have a stronger side wall structure in flexible filament than the square infill pattern below which has a tendency to "lozenge"

The advantage of a square infill pattern is that it allows the side wall to conform to the inside profile of an off the shelf shoe more readily without the insole buckling.

There are lots of different types of infill in various slicer packages so go forth and experiment, even MakerWare has Sharkfill and Catfill

Square Infill Pattern

Rotational Angle Increment Infill

Infill Rotation (Top) = 15°

There is an option in the CraftWare slicer to rotate each layer by a specified increment which has some interesting results. In the photos below, the difference between the two is remarkable and the properties of the insole changes too.

Infill Rotation on the left.

The insole on the left was sliced and printed with 5% "Rotational" infill. The one on the right was sliced more traditionally at 15% and the print times reflected this too.

It's another option to look out for.

One of the effects I enjoyed the result of was to use different coloured Filaflex throughout the print. With the Open Core mesh the colours really came though

Fancy a Bit of Colour?

Little strips of Filaflex Ready to Load

For those that want to know, I used the Recreus Extruder fitted to my Witbox with a PTFE feed tube modification. This is probably going to be another post to show it in full when I get time :

Here is a video of the Witbox printing one of the coloured heel inserts :

If you would like to have a practice at printing some simple insoles, messing around with the density values and patterns then download some examples from my Thingiverse page.

In the next article I will be writing about "Variable Density Meshes" and how to programme for these zones in Slic3r.

The disclaimer here is that I am only providing information, this is not medically certified and is not designed to be such, but hey in many parts of the world, people walk round on planks of wood or old car tyres so everything is builder beware. Go forth and MAKE!!!

Read Part 3 - Variable Density Meshes here.

]]><![CDATA[My Adventures with 3D Printed Insoles - Part 1/4]]>Fri, 12 Dec 2014 17:29:23 GMThttp://www.gyrobot.co.uk/blog/my-adventures-with-3d-printed-insoles-part-1-4Part 1 - Introduction

What I aim to show within the next few articles is how anyone without any qualifications, experience or knowledge in the podiatry field (that's me) can produce comfortable insoles using a completely free open source tool chain. There have been many 3D printed insoles before, certainly the more expensive multi-material printers have been showing off their exploits for a while now, however with this new democratised power of the maker movement, I would like to show and share some new found skills.

I am going to split this article into 3 further posts, each one focusing on 3 accumulative techniques that I have been experimenting with, each technique is not mutually exclusive, they can be mixed and matched with each other.

It can be assumed that I have produced all of my prints with Filaflex, the softest and most elastic material for FDM printers. If anyone wants an FDM printer recommendation for printing shoes/insoles then look no further than the Lewihe Sneaker printer, tested at 120mm/s printing of Filaflex and with a large build volume for shoes.

Here is a little overview of each post :

Part 1 - This Introduction
Part 2 - Open Core Mesh
Part 3 - Variable Density
Part 4 - Form Fitting

Open Core Mesh

Variable Density

The omission of the top and bottom layers of the print to expose the infill structure. This increases ventilation and can soften the feel of the print. It is also quicker to print and uses less material. Some nice infill patterns can finally take centre stage too.

With the later revisions of Slic3r and also Simplify 3D, there is now the ability to change print the properties for various discrete areas of the model to be printed. I did say this was an open source tool chain and I haven't used Simplify3D so my observations will be made using Slic3r. Slic3r can now use "Modifier Meshes" which are stl models imported onto the virtual build plate. Where these intersect your model, you can specify a whole new set of slicing parameter changes for this intersected zone. In the case of a variable density insole, it is the infill percentage that is modified.

Form Fitting

Another piece of the jigsaw, is how to create the surface of the insole to fit the form of your own foot. I will show you with the aid of my cheap 5 year old flat-bed document scanner how this surface is achieved.
It is difficult to describe in pictures just how flexible and comfortable this printed insole is.

What Next?

It's hard to say where this is going to go. There are many thousands of diabetic sufferers that have foot amputations due to ulceration, in far too many cases this is caused by ill fitting shoes and is preventable. This could provide a quick, low cost, and comfortable solution. What about a scanner and printer in every high street shoe shop or podiatrist clinic/medical centre, that would be a start.

It would be great to automate the various parts of the software tool chain into a wrapped up turnkey solution (free and open source of course), effectively to de-skill the process further. However this is outside of my individual skill set, although I am happy to collaborate further if anyone is interested.

The disclaimer here is that I am only providing information, this is not medically certified and is not designed to be such, but hey in many parts of the world, people walk round on planks of wood or old car tyres so everything is builder beware. Go forth and MAKE!!!

Read Part 2 - Open Core Mesh here.

]]><![CDATA[How to 3D Print with Flexible Filaments]]>Wed, 10 Dec 2014 16:53:17 GMThttp://www.gyrobot.co.uk/blog/how-to-3d-print-with-flexible-filamentsWhat is a Flexible Filament?

The saying, "You cannot push on a rope" remains as apt as ever when it comes to 3D printing with flexible filaments. Most printers will have some form of success straight out of the box with flexible filaments, but don't be greedy to begin with, turn the printing speed right down. Within this article I will try to explain as best as I can the problems, pitfalls to avoid and how to achieve success with pushing on a rope to achieve good print quality at acceptable speeds, in this case substituting the rope for flexible and elastic filaments for 3D printing such as Filaflex.

A flexible filament is generally a Polyurethane or Co-Polyester (not exclusively) Thermoplastic Elastomer (TPE) with a complex mix of additives/colourants to aid the 3D printing process. With good elastic properties and high co-efficient of friction means 3D printed designs in this filament expand the design possibilities. Effectively it is a soft stretchy grippy material.

A Bouncy Die

First Print

I know the feeling, you have just took delivery of your first reel of flexible filament, you load it into the printer, if you get as far as heating up the hot-end and actually getting to press the print button, then you have done well. My first time, the filament didn't even feed down into the hot-end successfully but came out of the side of the extruder from the hobbed shaft area making the usual spaghetti mess. I thought there would be a problem, because the first item on my hit list was a previously noted gap between the hobbed shaft and the entry into the hot end is a whopping 10mm.

Problem #1 - Unconstrained Filament

Extruder Problem

If a gap exists between the extruder drive gear (or hobbed shaft) and the entry hole of the hot end, then the filament can buckle and once the buckling process has started then it's too late and the filament will find a way out of the extruder sideways. If left unattended for a significant time, a flexible birds nest will have entombed your printer probably jamming a few axes motors in the process, which is not a wise move.

Even a small gap can have issues :

Extruder Problem

This gap needs to be closed up so the filament is constrained all the way into the hot end and cannot collapse sideways. You can see an example of this in the image above.

There are several ways to achieve this and your individual extruder may vary and it may/may not be possible to achieve on some printers.

1. Retrofit an extra support insert into the extruder.

An example of retrofit supports :

2. Reprint yourself a new extruder design with built in filament support.

There are many of these upgrades including several I have designed myself as part of "Gyrobot's Extruder Week":

And another alternative :

Mk8 Spring loaded Drive Block Replicator 1 / Duplicator 4 / FlashForge / CTC

Extruder Solution

Solution #1

Mk8 Spring loaded Drive Block Replicator 1 / Duplicator 4 / FlashForge / CTC

3. Purchase an off the shelf extruder and hot end which is flexible filament compatible, such as the Recreus Extruder which has achieved flexible printing speeds of 120mm/s (Due Nov 2014):

The Recreus Extruder

Problem #2 - Friction

Excessive filament friction is the enemy for flexible printing, it can cause more pressure on the drive gear consequently more filament feed problems can happen as a result. Both excessive friction before and after the drive gear can result in issues.

Solution #2

Before the Drive Gear

If possible, feed the flexible filament directly into the Extruder, without a using a feed tube. Ideally from above at a height of 12” or more from the printer, the higher the better. This will provide an unrestricted path, with gravity on your side. If you have to use a feed tube, make sure it is PTFE (low friction) rather than Silicon, fuel line or pneumatic pipe etc. Avoid any sharp turns or kinks in the tube, and allow the filament to enter the tube as straight as possible. It is unfortunate that that latest generation of some popular printers have the most convoluted path you could probably imagine.

After the Drive gear

Keep this length as short as possible to the hot end. Pushing a flexible filament into a hot end will result in compression and it’s diameter will increase as a result. Couple this with the thermal expansion of the filament inside the hot end and the pressure increases, miss feeds are the result. A Bowden style extruder are most problematic for this reason. I would suggest mounting the drive extruder above the printer rather than off to the side or rear, gravity will help in this case. Also for Bowden, as above, use a PTFE feed tube, and is also best with a 3mm diameter filament for fasted printing speeds, 1.75mm will work, again much slower though.
Insert a PTFE liner into the hot end barrel to reduce the thermal expansion problem :

Thermal Expansion Problem

Thermal Expansion Solution

The PTFE liner, reduces friction and acts a thermal break to insulate the filament until it enters the hot-end so the diameter doesn’t expand too early:

PTFE Liner Inside Barrel

Drilling out a hot-end barrel, lathe not necessary, a standard hand drill can be used instead

In the following photos, you can even see that the PTFE liner extends all the way up to the drive gear. This is the solution for an Orca 0.43 and my make of JackT's ORCA 0.43 Extruderderivate on Thingiverse

PTFE Tube Extending up to the Drive Gear

Filament Constrained Before Drive Gear too (Easier Filament Swapping)

WARNING : The maximum printing temperature for PTFE lined hot-end is about 265deg C

Problem #3 - Hotend Pressure

The last issue that we will address is a resulting property of fixing the previous two issues. If we have now have constrained filament and have reduced the friction, we can now theoretically feed filament into the hot end at a much faster feed-rate. However if the pressure builds up too much in the hot end then this can increase the torque on the motor and the filament can still buckle or compress sufficiently that the enlarged diameter of the compressed filament adds more friction to the side walls of the PTFE lined barrel, this again increases the pressure further and the downward spiral begins.

Solution #3

I have tried nozzles 0.2mm, 0.3mm, 0.4mm and 1mm and the 1mm nozzle certainly throws out the filament as you would expect, and you can ramp the speed up. A 1mm nozzle can actually still print pretty good quality prints, because the Z layer height can still be set low, say 0.3mm. The extrudate will be squashed flatter and you will loose detail on some complicated perimeters, but the Z resolution will still be good. A 0.4mm nozzle is a good compromise for XY quality and speed.
Cool the barrel. The smaller the melt zone and the closest it is to the nozzle the better. A fan blowing on the upper part of the barrel and over the general extruder area will keep this cool and help reduce any premature softening of the filament as it travels down the barrel. If the filament gets soft too soon then it can buckle easier and expand to grip the side walls of the PTFE liner.
Increase the temperature. The hotter the temperature of the hot end, the runnier the extrudate becomes therefore the flow is easier. 240deg is a common ceiling for this material before you may start noticing a bit of boiling occurring. If the filament starts to boil, the print will have a "foamy" look about it, which can actually still look quite good. Because the thermodynamics of various heater/thermistor hot-end combos are so different, it is difficult to tell you what exact temperature to print at. Try turning up the temperature and do some test prints, once you start noticing the material boiling, then back off 5°C

Final Word

Slicing

For me, my slicer of choice is Craftware by Craftunique. It seems to print flexible filament with the default settings, I just had to set the nozzle size (strangely called extrusion width) to my default 0.4 and reduce the speed accordingly, I also set the retract length to 4mm. It also has a great customisable support facility, which consists of pillars which can be added/removed/resized/auto generated before slicing, even placed on angles branching out like a tree. The support is easily removable, even with flexible filaments such as this Filaflex thumb below. It has super fast slicing, great graphics and is free, check it out :

CraftWare by CraftUnique

However, here are some good starting profiles using Filaflex for Cura or Slic3r.

If you are building your own printer or manufacturing printers, then consider the Recreus Extruder and Hotend combo which is due out at beginning of November 2014, It's a plug and play approach with many mounting holes on all sides so it is very adaptable. You will then be able to print with PLA, Filaflex and ABS at high speeds.

I would like to take this opportunity to thank Recreus for the use of some images to help me explain the problems and solutions.

There are also many solutions on Thingiverse for you to print off your own upgrade. Maybe it's just a case of tweaking your hot end or slicer settings.

I will leave you with a gallery of my prints which are either printed fully in Filaflex or incorporate Filaflex within their design.

]]>