These guidelines provide recommendations on design considerations for 3D models that will be used by touch readers. Given that good design is an art as well as a science, these guidelines cannot be entirely prescriptive. As explained in Principle 5, printing and touch testing is always necessary to ensure that your designs are effective and meaningful through touch.

Quick links to sections on this page:

• Design guidelines at a glance
• General principles
• Guidelines for specific types of models

Design guidelines at a glance

3D printing design element	Recommendations – minimum (min) and maximum (max)	Section
Base	Min 0.8mm thick. 1.5mm is recommended.	4.2
Stalks/stems/poles	Min 3mm diameter	4.2
Lines	Height min 0.5mm for faint lines; 1mm for definite lines. Width 1.1-3mm. Distance between lines min 2mm for double lines; min 5mm for separate lines.	1.5
Arrows	Raise the arrow head 0.5-1mm from the line	1.5
Heights	Min 0.5mm height for detection Min 1mm height for important features	1.3
Indents	Min 5mm wide for detection 0.5mm depth is sufficient. Indented print < 2mm wide.	1.3
Corners	Radius of 0.6 in TinkerCAD	1.10
Roads (on maps)	Indented Min 7mm wide 2cm wide recommended
Walls (on floorplans)	1-5mm height above the floor
Moving parts	0.3mm spacing between parts	2.3

General principles for the design of 3D models for touch readers

There are five main principles for the design of 3D models for touch readers. Each of these principles is described in more detail below.

1. Design for understanding
2. Design for inclusion
3. Design for engagement
4. Consider the 3D printing process
5. Iterative design based on touch testing

These general principles should be used alongside the guidelines for specific types of 3D models.

1. Design for understanding

Just as tactile graphics need to be a simplification of the print diagram rather than a direct replica, so too 3D models need to be a well-considered re-interpretation of the original object to best support understanding through touch. Touch differs from vision in that very fine details cannot be distinguished and an understanding of the whole must be built up through sequential active exploration of the parts through direct contact with the body – mainly the finger pads and perhaps also the fingers and palms. To allow for these considerations, objects designed for understanding through touch should be simplified to remove distracting tactile details, allow access with the fingers, and provide tactual distinction between elements.

In addition to following these design principles, understanding through touch requires accompanying information – context, a title and a detailed description – as described in the guidelines on Understanding 3D models through touch.

1.1 Simplify

Omit any redundant or distracting details that could otherwise confuse or distract the touch reader.

1.2 Spacing

Adequate spacing is required to distinguish between details and to allow space for the finger fingers to explore all surfaces of a 3D model.

As per tactile graphics guidelines, lines or similar details should be at least 2mm apart to allow distinction by touch.

Consider whether taller elements need to be spaced far enough apart to allow the fingers to feel the sides and base of the elements.

For two objects of similar height, a minimum of 5mm is required between them (Simonnet et al., 2018). If elements have a height difference of more than 1.5mm, then they should be separated by at least 15mm.

Consider the width of (adult or children’s) fingers when designing paths that should be followed or gaps that should be explored. Indented pathways work well for large maps but should be more than 7mm wide. If fingers must reach between two elements, they should be separated by:

• 1 finger width (20mm) if they are a similar height,
• 3 finger widths (60mm) if they differ in height by more than 1.5mm.

Tunnels and holes should be large enough for the fingers to enter. If this is not possible, consider outlining a hole so that it is more salient.

1.3 Heights

It is acceptable to distort heights and distances in order to enhance tactile distinction (Kaplan & Pyayt, 2022; Klopotowska & Magdziak, 2021).

Height differences need to be adequate so that elements can easily be perceived.

0.5mm is the minimum height for features that must be detectable but not prominent (Simonnet et al., 2018), however if this is less than the printer layer height the lines may be ignored. Therefore, a minimum height of 1mm is recommended for lines and other raised features.

Note that indents are much more difficult to perceive than raised features. At times, this can be used to your advantage, for example to insert print labels without disturbing the touch reading experience. Conversely, if you do wish to create indents that are important, they need to be wide enough for the finger pad to perceive. Alternatively, consider outlining a small but important indent with a raised line.

As described in section 7.5, exaggeration of the vertical scale is useful for topographic maps to enable detection of details in the landscape that would otherwise not be possible.

Keep in mind that tall objects can obstruct lateral hand movements for searching. Taller features need more space around them so that lower objects can be found and felt. Do not exceed ¾ finger length or 4cm in height for objects placed on a base with surrounding material, so that fingers can reach to the base.

1.4 Scale

Larger models are generally preferred (Watanabe & Sato, 2019). However, an exception is when objects can be replicated at actual size, which is easier to recognise than enlargements(Tu, Wu & Yeh, 2002).

If the 3D printed model differs in size from the original object, provide an indication of scale (Ania, 2020; Gesualdi & Griner, 2022). Some strategies may include:

• Adding ridge lines on multi-storey buildings to indicate each storey
• Adding a raised line to indicate the scale of an ant, 1cm, 1m or 100m, etc. as best suits the model
• Include an object of well-known size as part of the model, for example an ant, a dog, a person, a bus or a building
• If a small object has been enlarged for tactual exploration and understanding, print a companion copy that shows the actual size.

If very small details are important for understanding, consider providing a separate model, relief model or tactile graphic showing that section in more detail. For example, split the model into an overview to show the general shape and enlarged sections to illustrate the small details.

Given that 3D printer beds are usually quite small, it is possible to make larger 3D models by joining flat surfaces with glue (PLA) or acetone (ABS) or to incorporate joins into the design. Alternatively, multiple pieces can be printed separately and then affixed onto a large base. The overall size of a model should not exceed the length of the user’s arm (Sanches, Bueno & Okimoto, 2021). Also consider the space where the model is likely to be used.

1.5 Lines

Lines can be printed flat (with a square top) for a smooth finish and easy tracking.

0.5mm is the minimum width for detection of lines. However, 1-3mm thickness is recommended for reliable lines that can easily be detected.

As per guidance regarding heights (section 1.4), lines can be detected at 0.5mm but may not always be printed depending on printer settings. A minimum height of 1mm is therefore recommended.

Arrow heads can be printed 1mm higher than the lines so that they stand out. As with tactile graphics, arrow heads are easier to distinguish when presented as lines rather than a filled triangle. A length of 5mm is sufficient for the arrow head lines.

The distance between lines should be a minimum of 2mm for double lines; or a minimum of 5mm for separate lines.

1.6 Make the most important features easy to find

The most important or distinctive features of a model should be easy to find tactually.
For handheld models, this may mean making the most important features protrude.
For complex models such as anatomy, it may be helpful to print the model in parts that can come apart so the most important details can be explored individually.

For models that will be fixed to a base, the most important features should ideally be near the top of the model where they are easiest to find.

1.7 Don’t obscure the whole or obstruct hand movements

Enclosure using the whole hand(s) is an important method for gaining an understanding of a whole object through touch. Similarly, tracing along an edge or route is an important technique for gaining an understanding of the parts in relation to each other and the whole.
Unnecessarily tall or large features on a 3D model may obstruct such hand movements (Wang et al., 2022) and should therefore be minimised.

Permanent fixture of a model on a base restricts touch access to the lower parts of the model. If the whole shape is important for understanding, consider providing a detachable base plate or cradle rather than fixing the model to a base.

1.8 Use textures and material properties for meaning

A variety of textures or surfaces helps users to differentiate parts of the model. These textures can be integrated into the 3D printed model either as a surface pattern within the model itself, or through the addition of textured materials after printing (refer to our guidelines on Finishing).

• Shapes with a flat top (e.g. cylinders) are best for a smoother finish if printing on top of a shape; curved shapes (e.g. half sphere) are best if printing on the side of a shape.
• There is often no discernible difference between a grid versus random scattering of the same shape.

Cura slicing software provides the option of “fuzzy skin” to give a slightly textured feel on the sides of a model.  To access the fuzzy skin options, on the top toolbar go to Settings > Configure Setting Visibility and search for “fuzzy”. The settings can be adjusted for thickness and density. Fuzzy skin will be applied to the whole model. If you also want a flat texture on parts of the model, they should be oriented to the top or base of the print or you can combine two models. This feature is not suitable for models with small important details or braille, as they can be obscured by the texture. Check how it will appear after slicing in layer view.

1.9 Indicate orientation

When exploring 3D objects not attached to a base, touch readers tend to pick up the 3D model and rotate it in their hands. This is an important process for understanding the overall shape. However, it does mean that the orientation of the object can become confused.

When designing 3D models for touch reading it may be helpful to add a feature to indicate the bottom (flat), front (with braille label) or top (a dot or point) of a model that can be explained in accompanying information.

1.10 Stability for touch exploration

Tactile reading is a dynamic process, relying on movement of the hands and fingers over the surface of an object. For 3D printed pieces that can be combined in a variety of ways, ensure that once they are positioned, they can be kept steady while being touched.

Some suggested methods for providing stability:

• Create grid boards or frames in which 3D printed pieces can be placed. Laser cutting is a quick and easy technique for creating grid boards.
• Glue magnetic sheeting onto the back of 3D printed pieces then place them onto a white board or magnetic board.
• Glue non-slip matting onto the base of 3D models, or provide non-slip matting when the models are being used.

2. Design for inclusion

While touch reading should be the first consideration for accessible 3D models, it is helpful to incorporate features that will ensure the models can also be used by people who are sighted or have low vision.

2.1 Colour

If possible, print in colours that are meaningful in relation to the subject matter.

If a whole model will be printed in a single colour, avoid very dark colours and black, as these are more difficult to distinguish visually because the shadows are less visible.

2.2 Contrast

Use contrasting colours to highlight important features. Colour variation helps learners with low vision to distinguish and recognise components of a 3D printed object.
To print in two colours with a standard FDM printer, the design needs to be made in two parts (1 per colour) and the printer needs to be dual head.
It is also possible to stop a print part way and change the filament so that the base and top of the model are in contrasting colours.

Alternatively, colour can be added after printing using permanent markers, nail polish or automotive paint.

Recommended colour combinations for high contrast include:

• Yellow / Royal blue
• White on black
• Black on white
• Lime green on black
• Yellow on black
• Yellow on fire engine red (better for children with CVI)

Refer to Finishing for further advice on adding colour.

2.3 Supplement with print images

It may be helpful to supplement 3d models with print images, particularly if the model does not have high colour contrast and will be used by students with low vision. For example, Tactile Universe models are supplied with a reverse image on the base, so that the user can hold the model in front of them to look at the print while feeling the tactile model on the back.

2.4 Labelling

Incorporate print labels in addition to braille and, if possible, audio. Refer to Labelling for further guidance.

3. Design for engagement

3D printing offers an opportunity to create durable tactile learning tools with movement or multiple parts to provide active learning opportunities through interaction, play and reflection.

3.1 Manipulatives

3D prints are more engaging when they can be manipulated to create meaning. This may include:

• Spinners
• Multiple parts that can come apart and be put back together
• Multiple pieces that can be positioned in different ways to create your own meaning
• Hinges or other motions to explain physics concepts, show movement, or create flaps
• Bisections of models to show internal details, e.g. the internal cavities of the heart, or measurement lines inside a 3D shape

Refer to Moving Parts for tips on how to design these features.

3.2 Inserts for sound or weight

For engagement for younger children, you may like to consider adding objects inside a hollow 3D print so that it makes a noise when moved. Suggested objects include bells, stone, dried lentils or peppercorns. Larger and more expensive options include groan tubes or animal noise tins.

You may also choose to add inserts such as sand to increase the weight of a 3D printed object.

To insert an object inside a 3D print, include a space inside the model itself and then pause the printer when it is near the end of the space. 

4. Consider the 3D printing process

Design models with the printing process in mind. 

4.1 Avoid overhangs

Avoid overhanging areas as these will either fail to print or require support structures that result in a rough finish.

4.2 Design for strength

Avoid very thin structures that are likely to break. 

Bases should be a minimum of 0.8mm thick for strength, however beware of bending if you remove the print from the bed while it is still warm. For a more stable base, 1.5mm thickness is recommended for objects unlikely to be under strees, or 2mm thickness for stronger bases. Models with a very thick base use more plastic and take longer to print. 

Stalks or poles should be a minimum of 3mm thickness if short; or thicker if they are long. 

Also consider the direction of printing for thin parts. A 3D print is most likely to break along the horizontal printing plane.

When strength is of particular concern, also consider using a stronger printing filament such as ABS or tough PLA.

4.3 Design for economy

Reducing the thickness of a base (to a recommended minimum of 0.8mm) or eliminating the base entirely can greatly reduce the amount of filament and time required for printing.

As a further strategy to reduce printing time, braille labels can be printed separately – upright for smoothness and at a slow speed to reduce the chances of errors – allowing the main model to be printed more quickly and efficiently in a flat position.

When loading a model into your slicing software, orient the model to reduce the need for supports. This will help speed up the printing process, reduce plastic wastage and also minimise rough surfaces that are unpleasant to touch.

4.4 Design for re-use

Due to the printing time and space requirements of 3D prints, it is wise to design models so that they can be re-used for multiple teaching purposes. For example, if kept to the basics, a single topographic map can be augmented with Wikki stix to show rivers, highways or state/country borders and with Blu-tack or plasticine to show cities, towns or other features.

4.5 Safety considerations

Please note that 3D printed models are not considered food safe because they are porous. 

Corners and sharp points should be rounded for touch readers to avoid potential injury or harm and reduce the sharp plastic feel.  Rounding the corners of flat bases also makes it much easier to remove the model from the print bed. 

In Tinkercad, round edges and corners using the bevel and radius feature available for some shapes. A bevel or radius of 0.6 is sufficient. For other shapes that do not have these options, you can create your own shape to subtract from the corners.

To round corners in Onshape, Fusion360 and other more advanced 3D modelling software, use fillets and chamfers. 

To round corners in the code-based OpenSCAD 3D modelling software, use the hull() function with cylinders or spheres at each of the corners of the shape.

Prints should also be checked for sharp corners and points after printing, when they can be smoothed manually, as per the guidelines on Finishing.

5. Iterative design based on touch testing

Even when following guiding principles, design is an art as well as a science, and it is difficult to predict how something will feel when designing it visually on a computer screen. It is therefore important to touch test 3D printed designs and make adjustments according to the feedback. 

We recommend printing out small portions of a design, or at a smaller scale, to save on time and materials waste for initial touch testing of a new 3D model. 

Ideally, touch testing should be conducted by: 

• People who are congenitally blind to make sure that the concepts being used can be understood by them. 
• People who are novice touch readers to make sure that the model is tactually distinct for people who are unskilled in touch reading. 
• People either of a similar age to the target audience, or spanning a large range of ages if the model will be used by a broad population. This is because finger/hand size is an important consideration for design, and tactile acuity generally declines with age.

Design guidelines for specific types of 3D models

This section provides additional guidance on the design of specific types of 3D models for touch readers:

6. Models with moving parts
7. Maps
8. Icons
9. Art
10. Architecture

6. Models with moving parts

Manipulatives are 3D models with moving parts or multiple pieces that the user can arrange in relation to each other. 

Before printing parts that need to fit together or move, you must ensure your printer is well calibrated to be as dimensionally accurate as possible on the X and Y axis. Each filament’s properties can behave differently in the printing process, shrinking at different rates for different geometries (especially in printing holes). It is advised that you print calibration prints and adjust printer settings to compensate. See this video for how calibrate your printer when using Cura. This article further explains the process. 

6.1 Connections between pieces

Magnets provide an easy and pleasant way of joining pieces that are intended to be manipulated by the user, for example to add objects to a map or to join the cross-sections of a 3D shape.

To join two 3D printed pieces, we recommend using small neodymium magnets. Be careful to check that the magnets are positioned so that they attract rather than repel.

To enclose magnets inside a model:

1. Design a hole inside the model with a little extra height to allow inaccuracies.
2. Stop the print just before the hole is covered. You can either do this manually or by adjusting the g-code. In Cura, after slicing the model go to extensions > post processing > modify g-code. Select “add a script” and choose “pause at height” then enter the height for the top of the hole.
3. Insert some superglue and the magnet into the hole.
4. Wait for the superglue to act before resuming the print and watch carefully until the hole is covered to make sure the magnet does not lift out onto the metal print head(!)

It is also possible to design a hole at the base of the model and affix the model with superglue after printing, however this is not recommended for use with children under 5 years of age.

Avoid relying on joints or puzzle-piece connections for attaching parts, as printing parameters differ slightly from one printer and filament to the next, making it difficult to design joints that will fit snugly.

6.2 Moving parts

Strong tape such as duct tape can be used to join bisected halves of a model that needs to be unfolded to show the internal space. This strategy is explained in full by 3D4VIP.

Some models can be created with ‘print in place’ connections for spinning. The easiest method for designing such models is to find one that operates in the way you want then make modifications. However, ‘print in place’ hinges have a tendency to stick or break, so it is best to only use them where minimal force is likely to be applied.

When more strength is required, a better option for hinges is to 3D print a channel then insert a bolt or metal rod for greater strength (3D4VIP). It is also a good idea to increase infill when printing the parts with the bolt hole.

Another mechanism for movement is tracks that 3D printed pieces can be attached to using 3D printed rods or a nut and bolt.

7. Maps

7.1 Spacing

Consider widening streets and gaps between buildings so that the fingers can feel the sides of buildings and easily trace along routes. 7mm is sufficient but wider streets may be better.

7.2 Heights

The height of buildings and walls should represent their relative height but the vertical scale can differ from the horizontal scale. Heights should be low enough for the fingers to easily reach the base.

On street maps, consider adding ridge lines on multi-storey buildings to indicate each storey. If space permits for roads and paths to be wide enough for the finger tip, they should be lower than surrounding features to reflect the act of stepping down from the kerb.

On indoor maps or floorplans, walls should be higher than floors and pathways so that they can be understood as walls. A height difference of 1-3mm between pathways and walls is sufficient (Kaplan & Pyayt, 2022) and up to 5mm is acceptable.

7.3 Realism

Where possible, indicate features with (simplified) realism rather than abstract symbols. For example, stairs and trees can be represented realistically and are easy to understand without reference to a key. Refer to section 8 Icons for further suggestions.

7.5 Topographic maps

Topographic maps of a chosen region can easily be created using the free Terrain2STL website.

Super-elevation, i.e. an exaggeration of the vertical scale, is useful for topographic maps to enable detection of details in the landscape that would otherwise not be possible. As a general guideline, 3D4VIP recommends a super-elevation factor of 3, whereby the z values (heights) are represented three times as high as the scale in the horizontal plane would actually require. Accompanying information should explain the super-elevation to the touch reader.

8. Icons

3D icons for maps can be recognised more easily than 2.5D icons, reducing the need to rely on a key and making tactile maps easier to understand (Holloway, Butler & Marriott, 2023).

A selection of pre-tested icons for maps of shopping districts, parks and playgrounds are available for free download from Thingiverse at www.thingiverse.com/thing:5841775. These icons should be printed at a minimum size of 2cm³. If space permits, larger icons are generally easier to recognise.

The existing icons are provided on a circular base so that they can be securely fixed to the map base. Alternatively, they can be embedded into a map model and printed all in one piece. Icons must be spaced apart to allow adequate access for tactile exploration from all sides.

Further icons can be designed based on the following guidelines:

• Use objects that are a distinctive shape. Objects that are regular shapes such as rectangular prisms and rounded objects cannot easily be distinguished unless they have very distinctive additional features.
• Common objects that are hand-held are ideal for icon representations because they are most likely to be familiar to people who are congenitally blind.
• The most important and distinctive features should be near the top of the icon, where they can be more easily accessed by touch. It may be necessary to position the object so that the distinctive features can be felt.
• Simplify the object to its bare essentials removing any extraneous details
• Exaggerate features that are distinctive. For example, a dog or cat’s ears should be exaggerated to assist with identification. However, this exaggeration of features should not be based on stylistic visual conventions that would be unfamiliar to a touch reader who has been blind since birth.
• Flatter icons are more difficult for novice touch readers to recognise than those that are more 3-dimensional. Consider rotating the object to an upright position so that it is easier to feel.
• Corners and edges should be square or round in accordance with the original object to assist with interpretation. Often this means square corners for man-made objects and rounded corners for organic objects.

9. Art

9.1 Sculptures and museum objects

In general, for the representation of sculptures, 3D models are preferred to tactile reliefs, which are in turn preferred over tactile graphics (Fuller & Watkins, 2011).

A large range of 3D models for famous artworks are available through the Scan The World project. However, when selecting models from this collection keep in mind that the original artworks were not designed for 3D printing, so many of the models may be too complex for FDM printing, require a lot of supports, or include delicate structures.

Materiality is important to make objects like museum artefacts authentic. For tips on adjusting the surface feel and weight of a 3D print, refer to our guidelines on Slicing and Finishing. However, keep in mind that 3D printing may not be the best technique to represent objects with very different material properties. For example, to represent a jellyfish it would be much better to 3D print a mould and then fill it with jelly or a similarly soft, wobbly substance.

“All models are wrong, but some are useful” – George E. P. Box

9.2 Paintings

Paintings are much more difficult to render successfully as 3D prints because the original format is flat, they often rely heavily on colour and the scene being depicted may be very complex.

The first option for representing a painting using 3D printing is to create a relief model. A successful depiction will simplify down to the most important details. Scenery in the background can either be omitted or shown very flat, with foreground objects much more raised.

A second option for representing a painting using 3D printing is to create a tactile graphic showing the basic composition then accompany this with 3D printed models (and other artefacts) of the most important elements within the painting.

Lastly, an abstract or semi-abstract painting can potentially be represented by converting greyscale to heights in a relief model. This method may be suitable for showing broad brush strokes and abstract shapes but should NOT be used for representational artworks.

9.3 Accompanying information

Perhaps even more so than other 3D artefacts, 3D printed models of artworks need to be accompanied by descriptions and other modalities in order to convey not just the subject matter but also the colour, lighting, shadow, materials, scale, context, aesthetics and emotional response. Art concepts may also need to be explained.

As an example, a 3D model of a sculpture may be supplemented with a sample of the type of stone from which the original was carved, or a painting of sheep may be accompanied by a sample of unwashed wool and the sound of bleating sheep.

10. Architecture

When representing buildings and architectural models using 3D printing, adjust the level of detail according to the scale and purpose of the model.

For a small building on a large map, simply provide the basic shape to convey the shape, scale and position of the building.

For a tall building on a large map, provide the basic shape with an indication of the number of storeys. This can be conveyed with indented windows on a smaller building, or with ridges for each level on a skyscraper.

When the building and its architecture is the main topic of a 3D model, identify the key architectural features for inclusion and ensure that they are able to be felt tactually. This may require some distortion of scale, for example exaggerating offsets, widening columns and chimneys (Barvir, Brus & Vondrakova, 2021), bringing doorways forward or adding a raised outline around windows and other indented features. Remove ornate features that will cause confusion or be easily broken.

To show internal details of a building, consider making the roof and upper levels removable. Internal walls do not need to be full height, so that the finger can easily reach to the floor where braille labels and travel routes may be marked. The removable parts must be designed so that they are easy to remove and return to the correct position.

Ornate architectural details are best conveyed as enlarged tactile graphics or 3D printed relief models rather than trying to include them on a 3D model.

---

Last updated: September 22, 2023 at 9:34 am