Braille Computer

A low cost Raspberry Pi powered all in one Braille computer with a full page display.

Similar projects worth following
A traditional Braille terminal links to a PC and provided a single line of text. These devices cost upwards of £1800 plus the cost of a computer and require training putting them out of reach of most blind people. In 2007 it was found that only 10% of blind children could read Braille, due to access in education compared to 50% in the 1960's. My solution it to build a low cost all in one system which can also teach, much like the Khan Academy.

The reason a Braille terminal is so expensive is the cells which push each dot on demand. What I will do is use a cnc type mechanism to push cheap resettable 'pins' in order to provide a full page of information. A complete computer should hopefully come in under £200.

Competing Products - Best prices from RNIB website

Next Generation Perkins brailler, raspberry

Basic Typewriter for A4 paper.


Seika mini braille display

Braille terminal for controlling portable computers with a 16 cell wide display.


Braille Sense U2 portable braille notetaker

"Use the pre-installed software such as the word processor, email and web browser or link to another device such as an iPhone and use as a 32 cell braille display. You can even run several programmes at the same time and easily switch between tasks."


Proposed Specs

30 cell wide by 20 row high A4 portrait display.

Each cell will be 8 bit to allow for international and enhanced standards, and also an underline for hyperlinks.

Display will be single point touch sensitive, with link buttons to the side of the display.

An HDMI port will be accessible for the benefit of sighted tutors and parents.

Built in speakers and microphone.

An indented cover for the display which can sandwich an A4 sheet of paper to 'print' a page.

'Pixel' mechanism

Truth table


The top pen demonstrates the off position, the middle the on, and the bottom shows the reset. This acts as a mechanical SR latch flip-flop. The advantage of these pens is that it is unnecessary to remember the last state, as the set and reset are independent with the reset being on the side (and a pack of 10 cost me only £1 :). A small demonstration video will be added here when i have had enough time with my dremel, superglue and polymorph.

  • 60 × mini steppers dslr style
  • 1 × large servo
  • 1 × raspberry pi 2
  • 1 × teensy lc for hid interface, touch sensor
  • 1 × atmega Braille display controler

  • Simplified Mechanism??

    haydn jones06/11/2015 at 20:30 0 comments

    This awfully rendered animation that demonstrates my simplified mechanism. So what i need to make is three boards with 4800 holes (apart from mounting) and 4800 pins made. Any ideas? :) I can reduce the design down to 20x12 6 dot characters and reduce the number to 1440, which I think is still a good size display.

  • Thoughts on blog feedback

    haydn jones06/02/2015 at 11:20 0 comments

    So @Benchoff posted a well written article about the Braille computer and the matching keyboard, articulating my ideas much better than I could myself. After the initial panic of seeing such a messy and early version of my project being seen by people who's work I've been following for years, I started reading the feedback and its been really helpful. To clarify, my design has 60 servos placed along the x axis, which is pulled along the y axis by another servo, like a photocopier or desktop scanner in movement.

    Here are some of the ideas and my thoughts on how I should progress.

    I had planned of using SG90 Tower Pro servos as they are cheap and everyone uses them, until @Robert Guyser showed me these. They are camera lens motors, and some are just 6mm in diameter and light at less than $1 each. I wont have to worry about clumsy mechanisms in order to reach the dots.

    Electromagnetic displays were mentioned, but i have no idea how to deal with cross chatter at such a fine pitch. Single line displays use a bulky servo array.

    A few people mentioned the pin art toys (block of metal pins people in the 90s used to copy their faces) if the pins were ridged a system similar to a zip tie could hold them in place, with a global per line reset.

    [Rob] suggested electric stimulation in place of dots, don't think i would like to use it during a thunderstorm :)

    [W] posted this...

    "If your goal is a simple/cheap output device, why not ignore Braille entirely? Since people can use a chorded keyboard for input, why not use a chorded output? Think of 8 pins/buttons in the same pattern as the chorded keyboard, each lifted by a solenoid. Use the exact same patterns that you type on the chorded keyboard with.

    You'd need a rate control, but that could be handled by a foot pedal or something. I agree it's nowhere near as convenient to use compared to a printed braille page, but it should be an order of magnitude cheaper than a braille printer and therefore more widely accessible.

    If you were mechanically super-clever, the output device could also act as the keyboard."

    This last one is quite interesting. Beyond the scope of this project, but it has loads of potential for mobile.

    Thank you to everyone who left feedback, it was all appreciated :)

  • Slow Progress is Progress

    haydn jones05/27/2015 at 21:11 0 comments

    The keyboard is coming along nicely (see #8-bit binary/hex/braille keyboard) and I finally found a flatbed scanner for my screen/enclosure. I have to design the actual Braille points in cad soon and use my shapeways credit. Because i seed thousands of small pieces (but only 4 different shapes) I am planning on getting a metal mould and trying my hand at injection moulding, but I will post on here before I go ahead.

    Dog for scale

  • Keyboard module ready

    haydn jones05/06/2015 at 19:56 0 comments

    Just breadboarded my keyboard module #8-bit binary/hex/braille keyboard so I now have a starting point for the physical interface.

View all 4 project logs

Enjoy this project?



Dylan Brophy wrote 03/29/2017 at 00:34 point

I love this!  It's an excellent problem to solve. That's crazy that 40% less blind people can read braille.  Nice that someone is trying to solve this problem.  Also, the idea of a mechanical output for text is very interesting to me.

  Are you sure? yes | no

PAUL DSOUZA wrote 02/21/2016 at 16:31 point

I have posted many more videos of the different approaches I tried over the years... look at my channel.  It might save you time not making the same mistakes I made!

  Are you sure? yes | no

PAUL DSOUZA wrote 02/21/2016 at 16:27 point

Hello,  I must commend you on the excellent work you have done! We seem to be following the same path.  These are some links to some videos where I demonstrate some Braille Display mechanisms - including a similar common latch mechanism. The actuators are different. I also have a single line version that works quite well!  The refresh is mechanical - lower cost.

  Are you sure? yes | no

jeremy.darling wrote 02/09/2016 at 18:30 point

Any updates or word on this?  I recently got involved with a family with a blind child and I must admit that I was taken aback by the cost of these machines.  I immediately started thinking of how I could "hack" something together, and then ran across your HAD page.  Looks like you have already done a lot of ground work, and I'd love to help where I can.  I might start something of my own, but as I tend to get half way through and then run out of time/energy/money/something else I really would rather team up with someone.

  Are you sure? yes | no

b4ux1t3 wrote 09/29/2015 at 13:17 point

This is a really cool project. I had a couple thoughts, based on my observations working with my mother's visually impaired students.

1. A lot of blind children get *really* good at typing on normal keyboards. One kid I am working with is a faster typist than *I* am, and I've been typing for almost 20 years longer than he has. It might be worth it to forget adding a keyboard in favor of building the display (at least for the short term), while allowing for a user to plug in an external keyboard of their own (with which they will likely be much more familiar)

2. Unless I'm misunderstanding something (have read through the whole page, though mostly very quickly) it looks like you want to be able to detect touches on the display. This is a great idea, but one I feel is less important than you might think. While one would think a blind person could hit a hyperlink on their display (maybe by pressing a button on the side and then tapping lightly, or double tapping, the link text), in practice, it might be better to just have some functionality to cycle through hyperlinks in web documents (or specific sections within documents) and then select them to follow the link. Though, maybe some combination thereof would be better, I'm not sure.

I don't know how much access you have to people who are visually impaired, so I'd be happy to ask the ones I know any questions you might have. I assume, however, that you're doing this *specifically* because you know someone who is visually impaired, so maybe I can just help by providing a larger sample size.

Anyway, this got a bit long. This is a great project! If you've made any progress in the past four months, let us know!

  Are you sure? yes | no

matseng wrote 07/16/2015 at 09:53 point

It's impressive that blind persons can read braille, I have tried to feel the difference between the bigass braille dots that usually are on the buttons in a lift - but I can't really feel anything at all.

But are the blind equally sensitive to small heated regions? I.E could they read braille dots that are made out of warm and cold regions instead of raised dots? Is so one could build a solid-state non-moving braille display....

  Are you sure? yes | no

Sergio Costas wrote 06/06/2015 at 20:16 point

If you accept suggestions, I strongly recommend a different approach: just use a mechanism like the Zuse's Z1 mechanical memory. The idea would be to use a basic cell like this one:

The red dots are just cylinders that can be in one of both positions. Over each "bit" will be a "dot" that will be up when the cylinder is in one position or down when the cylinder is in the other position. This can be done easily if the dot is like a wedge.

The interesting thing about this idea is that this cell can be repeated in 2D and each one connected to its neighbors; this allows to control X*Y dots with just X+Y solenoids (or whatever you want to use). Even more: the side with more bars can be multiplexed, in order to use log2 solenoids and reducing further its count.

Also, the cells are built with just metal sheets, which are very cheap and easy to create.

  Are you sure? yes | no

Trevor Johansen Aase wrote 06/02/2015 at 21:35 point

Great idea! I have some ideas to throw into the mix:

Use a 2x3 array of micro solenoids on a sled (like a CD tray ejecting) moved to each character via a hobby servo. Put a light sensing diode in the middle of each character so the sled knows that char has been read and can race behind updating the next bit of data. That way refresh rate would not really matter. Stretch a piece of latex or something similar over top of everything to give it a nice smooth yet defined surface.

  Are you sure? yes | no

haydn jones wrote 06/02/2015 at 21:57 point

Hi Trevor

Thanks for the feedback, I'm actually useing the 8 dot Braille standard as it is better for science, but I get your point. Your idea is interesting, but maybe a bit to complex for me. I am using single bar of steppers (I couldn't find solenoids cheap and small enough) and setting them as an entire row, so 4 steps is an entire line of text, with each line resettable on its own. It would be much better if I could reset each character on its own, but the mechanics would be too fragile. If I can prove the concept, I can refine it in later versions.

  Are you sure? yes | no

IT-Wizard wrote 06/02/2015 at 16:14 point

Great idea. It is true that disabled people does not have an easy life.

As I understand, you want to use the same mechanism as the pen. It means a spring for every digit.

May I suggest to use a system like in the computer keyboard as spring. Small rubber stuff.

  Are you sure? yes | no

haydn jones wrote 06/02/2015 at 17:34 point

I hadn't considered a membrane, I will look at the feasibility for my scale.

  Are you sure? yes | no

IT-Wizard wrote 06/02/2015 at 18:23 point

I understand, but I think how to bring it to big and cheap... :o)

Keep on the your prototype.

  Are you sure? yes | no

haydn jones wrote 06/02/2015 at 18:28 point

I meant component size, would it fall apart because its too small? I've had so much feedback in the last day or so I'm almost having to completely redesign before I have even started :) Thanks for the support, always looking for new ideas.

  Are you sure? yes | no

Shawn Kirsch wrote 06/01/2015 at 22:17 point

will be cool to see where this goes.

  Are you sure? yes | no

Ahmed Hefnawi (Volta) wrote 05/05/2015 at 20:28 point

Brilliant Idea!

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates