Hackaday Prize Entry: A Braille Computer

As with all devices meant for a very small percentage of the population, computing equipment for the blind is very, very expensive. A Braille typewriter – a relatively simple machine that puts dots on a piece of paper – costs about $700 USD. Need a Braille interface for a computer? You can buy a 16-cell wide Braille output for $1600, and high-end models with an integrated keyboard go up to $5000.

For his Hackaday Prize entry, [Haydn Jones] is building a simpler and cheaper Braille computer. It’s not just a single line of text at a time; this computer will have a display that will output an entire page of Braille at a time.

The current solutions for a computer to Braille interface use small electromechanical cells for each character. That’s six individual pins for each character, multiplied by the number of cells on the display. Doing a full-page display with this type of mechanism, but [Haydn] has another idea. Instead of controlling each pin individually, all of the pins on the display will be controlled by a CNC-like mechanism. The pins themselves will be mechanical SR latches, better known as the mechanism in a ball point pen.

A display is only half of the IO of a computer, and for the input portion of his build, [Haydn] is also building a Braille keyboard. This doubles as a binary or hexadecimal keyboard, but the idea is very similar to a proper chorded Braille keyboard. It’s a simple enough build; just a few key switches and a microcontroller.

The 2015 Hackaday Prize is sponsored by:

62 thoughts on “Hackaday Prize Entry: A Braille Computer

  1. Ambitious.. The ‘screen redraw time’ sounds really slow.. Like profoundly slow.. maybe 100CPM…

    This is a very difficult problem to solve – the braille readout problem.

      1. A braille readout is difficult.

        Let me know if you need CAD help – I am (without modesty) a skilled SolidWorks expert and can help you prepare designs for 3D printing.

  2. For the readout, how about this idea?

    The “display” is made of hundreds of bits of memory metal. For each screen refresh, a mechanism flexes each dot and whatever heating current is required is applied. Then using X/Y type addressing, dots are heated to return back to their default flat state.

    The advantage is there is no mechanism which needs to scan: the mechanism uniformly addresses all dots at once, so alignment can be fixed and therefore simple. The addressing is done electrically, which is a lot faster and uses well-known addressing techniques, a bit like core memory.

    1. I think reliability and speed are an issue with this.. There are similar methods that actually are used, and they are accompanied by somewhat complex locking mechanisms often.

    2. One ciuld just use loop of paper od thin sheet of metal between two rollers and place printing mechanism on the bottom side of the loop and table under the top side. Then add movable flattening roller before the printing head. This way one can have persistent display of any size with ability to delete only selected lines. It needs only one six or eight cell printing head which would be cheaper than Braille output. the only downside is the noise. Braille printers are quite loud. A friend of mine had one that sounded like machine gun.

      1. I have recently seen a concept like this that uses a sheet of plastic, billing itself as a ‘3d printer for braille’.

        potential for low cost is great, but usability as a digital display\interface is severely reduced.

      1. The heat is basically a deal breaker. I worked on braille readouts for my college senior thesis design project.

        Commercial arrays mostly use tiny solenoids, which also have issues – with current drivers, locking mechanisms, etc.

        Memory wire still seems like a good potential approach, but the motion is very small, and as you say it makes a lot of heat.

        One potential, especially just for a prototype is these ‘linear stepper motors’ – ive seen packages of 20qty of them for like $12 from asia. These do not cut down complexity though, and likely have their own distracting drawbacks.

        There is printers for braille – they use sheets of plastic and a single pin dot matrix style printer. This is another interesting area to check into.

        Phase change materials – like liquids that turn solid, etc. with a current are a good place to look as well.

  3. Interesting thing. I had the same idea (a braille full page), but decided that it must be even simpler. In fact, I had the idea when I saw the mechanical memory from the Zuse’s Z1. I think that it must be a system more like that, based on mechanical multiplexers.

    In fact… I think I’ll try to develop that.

    1. Having just glanced at the mechanical memory of the Z1, i can affirm that this looks VERY similar to the traditional approach to creating a refreshable braille display – extremely similar mechanism. Good intuition.

  4. Considering you can only read one line of braille at a time it seems kind of redundant. Your eye can easily scan a visual page to land on what you want to read and scrolling up and down is an additional physical movement that is unwelcome as it interrupts the flow.

    Reading braille however is purely a physical movement and trying to scan a page to find what you are looking for requires that you have to guess where what you are looking for is. For example if you want to skip a paragraph you would have to move your reading finger down several lines, maybe looking for a gap in the text.

    So therefore surely the best option would be to have a single line that features capacitive touch so that the computer knows the position of the readers finger. As the finger reaches the end of the line a tactile feedback could be given to indicate the end of the line. The reader could then perform a number of actions, for example press to re-read the line or return their finger to the start for the next line. Perhaps a double tap would skip to the next paragraph.

    1. Single line braille readouts are more common than full page ones, actually.

      They are far less expensive. Blind people often have limited resources.

      Sadly (or maybe not, i dont know) braille is basically being replaced by text to speech technologies. On one hand it sucks, but on the other hand its far cheaper and less mechanically intense.

      1. The problem is, even if blind people could afford a $1600 terminal plus pc, they can not use it without education. Most blind people are not taught how any more. I want this computer to mirror on to a monitor so a parent/tutor can intuitively guide the student. Braille literacy greatly increases employment prospects, and is essential for sciences/mathematics. I really hope I can get this right.

        1. My idea in college was for the readout to be a playstation peripheral. it was my thought that this opened it up to a huge market, with a ‘computer’ with a massively simplified physical interface – just a power and reset switch and a cd-rom slot.

  5. Just a thought– While the pen mechanism is ubiquitous/affordable, I wonder how one might be able to facilitate the ‘reset’ action.

    Instead, does anyone remember these ‘toys’ that were popular in the 90’s for awhile ? [Example: http://goo.gl/SYbcgk%5D Really pretty cheap considering in some, at least hypothetical sense, it already contains the perfect layout.

    So the idea then would be, beneath the pin base to create an electromagnetic mesh network, just like a ‘keyboard’, but here in reverse– Or here you are not taking in the input of a key, but rather outputting it.

    Granted, electromagnetic is not my personal area of specialty, but my ‘first order’ thought as to how this might work is like this (Also, excepting I do not have one of these toys myself, I am not sure their metallic properties or what they are made of):

    An inexpensive, permanent magnet underneath the base (basically to provide the attractive/repulsive force and keep all else ‘settled down’). Then, individually, one at a time, carefully remove the pins, do your wire wrap, and heat shrink the result.

    Micro-controller thus dictates ‘reverse polarity’, sending the pin up. And yes, I worry a bit about cross field forces kind of flummoxing the whole thing up, but perhaps someone out there with more experience in EM could speak better to the specifics of number of wraps, relationship of amperage, and the size of the force field thus created (as when you are tripping one pin, well okay, but several in a small space, side by side, at least in my experience once electronics interacts with the physical world, there are a lot more ‘unexpected effects’ you have to start thinking about).

    The ‘cool’ thing about this is, especially, if there are readers out there, ‘Moore’s law’ is actually very, very slow, or yes, today we can switch an exponential number of transistors in– But just take a moment to think of someone, everyday, that can read by ‘touch’.

    I am more in ‘awe’;

      1. I guess that he’s talking about those grids of pins, that you could put on your face or hand or whatever and make spooky-looking “sculptures”. No idea what they’re called, pin grid toy might Google it. Ahh…. it did. It’s just called “pin art”. I think it would probably be more frustrating trying to electromagnetise this than to start from scratch. Putting pins in a grid is probably the least difficult part of a full-screen braille display.

        I can’t help thinking some little robot, like a LOGO turtle, if you remember, would be good. Doesn’t necessarily need pins. Just some way of leaving a feelable trail. Rather than hundrds of little pins. No idea how, like. Maybe there are bistable materials that might help.

        Or maybe core memory, or flip-dot displays have something to offer. People used all sorts of wierd displays before fully electronic ones got cheap.

  6. Since a blind person cannot see, it seems a waste to have a full page of text (or even a full line) that is unobserved. Especially if it takes 30 seconds to generate a full page. They can read, at most, 4 characters at a time (using four fingers).

    A single character is really all that is needed with some type of intuitive interface to give the sensation of running a finger across it and possibly even gesture recognition to determine what the reader is wanting to do. Making a pad that they rest the palm in with 4 characters (one at each finger) could simulate the sensation of scrolling across a page and the thumb could even be used for interacting with it.

    1. Imagine reading Braille to be like tunnel vision, except you can find 8 points at once. The current systems, apart from paper books are like scrolling through an infinite ticker tape or a 16 character terminal. You can’t have graphs, charts, tables, complete math formulas and you can forget about programming. Being able to find the first word of each line could (hopefully) open up a lot of opportunities.

      1. Maybe a lo-res representation of the display, with each character just being a dot. Then once you’ve felt out the page’s structure, and guessed where you’re going, the display detects your finger and shows the actual letters on a separate 16-character display.

        Still, even 80 columns by 25 is 2000 characters, or dots. It could be slower than the main Braille reader part, since it’s only refreshed once per page.

  7. Have on idea on navigation thru text. Since it is difficult to scan using the physical motion of fingertips across the display perhaps a different approach from old software.

    I used a dumb terminal editor with DEC minicomputer systems. The keyboard included a numeric pad to the right. Used a text editor (TECO?) which configured the numeric pad into navigation and editing buttons. You could advance/reverse (first key press) by word/sentence/paragraph (second key press) your position in the document being edited.

    So you could Select -> Advance -> Row, Row, Row (couldn’t resist) -> Copy. Navigate elsewhere the Paste. Pretty efficient without a pointing device.

    Perhaps combine a few navigation keys used by one hand (advance -> paragraph) to refresh a row or two of braille text for the other hand (you are at this text).

  8. I can see a full page braille tool having tremendous use. Think of the possibility of a zoom… In a word processor, zoom out gives a textural representation of the paragraph structure. Farther out, the chapter structure. Zoom in to get the readable text. Scroll wheels to the sides, top, bottom for zoom and scroll. This would require moderate speed (maybe a couple seconds max for a full sheet) and would be enhanced by the display being touch sensitive (zoom in on touched point, for example).

    I can think of a few mechanisms that could do this fairly fast, and likely be quiet enough for the home environment, possibly at higher density than needed strictly for the braille text, to allow for such things as graphs, though the ones seem most practical are going to involve magnets to pick state (think a latching relay) and this could be an issue.

    1. It kind of is – except a tiny flip dot display is a very intense mechanism without much simplicity to it.

      All the ‘new’ ideas for braille readouts are either r&d centric experimental approaches, or things like ‘3D printers’ that work specifically to create pages of braille, etc.

    1. Great example\demo,

      I think maybe if each element was individual – with a rubber bladder or a plunger or something, this could have promise.

      I suppose to hold a whole sheet stable, this would take amps of current.

  9. Great idea!

    For handling the concern people have about slow refresh, can you optimize the print head path to what people would commonly use? Most people might read the top line first, then maybe scan the first word of each of the first few rows. If you set those pins first, they could start reading before the page is fully rendered.

  10. What are the size requirements for each “pixel” and what is the spacing between them?

    I was thinking something like patches of fine iron filings between 2 membranes. the top membrane being soft, and the bottom harder, with tiny electromagnets under the bottom layer. When the magnets are off the filings lay flat, but when the electromagnet is on, the filings bunch together in a ball and push up on the top membrane?

    I donno…just throwing an idea out there.

    1. There is some common size – but they are roughly 1cm x .75xm and the dots themselves are very small – picture a pen tip pushed into pliable plastic from the back.

    2. What about something like an etch-a-sketch, a moving stylus that affects a bistable screen? Or more like, Magna-Doodle, the toy with the iron-filing display, where the screen has just enough magnetism to keep the filings stuck til they’re mechanically wiped. The key, obviously, is making it feelable. But using a stylus means just one mechanism instead of thousands, I think that’s the way, especially for cheapness, reliability, and nobody going insane trying to build one at home.

      What about wax, maybe? Inscribing it mechanically, then maybe a little heat and a sortof squeegee mechanism to flatten it, without having it go full molten. It would beep as a warning for when it’s going to heat up. Perhaps just the wiper blade would need to be hot.

      1. I’ve just had an idea inspired by the bio-hackers who were mentioned here in the past… What about magnets on the person’s fingers? Strapped on rather than implanted, although maybe if it’s an advantage having tiny magnets in the fingertip, some would do it.

        It takes advantage of our sensitive sense of touch. So you’d need a grid of electromagnets. It would probably be fairly coarse, but just to give a representation of a page’s structure, where the paragraphs and lines are, as well as simple pictures, might be good enough. Take the magnets out of the display, and take out the moving parts too!

          1. Bit disappointed nobody else thought so, I thought it was genius, if I do say so myself. Option might be lots of little magnets attached to a finger-glove, or even some magnetic granules glued to the finger. As an experiment. And then if you can find some hardcore piercer dude who happens to be blind, or just fancies another sense, having tiny beads or grains of magnet embedded in skin.

            The idea would be to use as small magnets as possible, spread out as close to the skin as possible, to keep the tactile resolution high, certainly since Braille can be read in quite a small font, the sensitivity is there in the fingers. From experience with playing with magnets a repulsive field is quite easy and strong to feel the shape of, it pushes back against you.

            So yep, tiny wee magnets on or in fingers, are there many blind / partially-sighted body modders? I’d assume from purely guessing, that the point of mods and tattoos would be lost on blind people generally, but who knows? If a few start off doing it and it works really well, that may well convince others. Magnetic fingertips wouldn’t be much of a handicap in general life, compared to the advantage of a cheap and good display.

            Hm maybe if you’re a real hardcore blind person, Braille tattoos! These genius ideas are flowing pretty freely here! Maybe 3D printed, then a mould made and cast the thing in medical silicone? Or punch bumps into a small sheet of surgical metal, with lots of holes in the plate for the flesh to knit through. Would be a hell of a fad and keep the media entertained for half a week.

            I wonder how the existing magnet-fingers body-mod chaps are getting on? At least a few were geeks, what have they come up with? Off to look…

          2. Or put the tiny neo magnets in a thick plastic sheet – pretty much like a bubble wrap where each of the bubbles contains a magnet. Populate a PCB with similar sized unshielded SMT inductors as electromagnets that you can address. Use beef up LED matrix drivers to modulate current to the inductors.

  11. I still maintain that a full page is too much. Who says the display needs to stay in one place? Using the example above of a cradle that you rest your fingers in with braille character display under the fingers… just make it like a mouse and roll it around the virtual page.

    1. Blind people can read braille more quickly than it can be refreshed – the larger displays are greatly preferred by the actual users for the same reasons sighted people like full pages – blind people can also remember where on a page they are, quickly scan lines, etc. Noise is another concern that wasnt mentioned here yet, but traditional braille readouts are really noisy! Updating a page at a time will be less noise than a line at a time.

      1. ??

        A) All things you can do with a mouse based display (did you forget that they can’t see and need to feel for that spot or remember where to move their hands anyway?)
        B) Noise is a matter of mechanism used. Not inherent. And having the fingers resting directly on the characters would dampen the sound considerably.

        Not my project, though.

        1. Mine neither! I dont understand your comment exactly.

          Having observced people actually using braille readouts, they hold the thing like a tablet, or place it on the table in front of them, there are wheels on the device (or buttons) to scroll, like any person, they can remember where on a line they are, and ‘sight read’ quickly to find their place, and so on.

          There is a physical aspect to using a braille readout, and unfortunately getting all the little pins to cooperate, up to now, has resulted in fairly noisy mechanisms.

          A ‘full page’ is not a known quantity – displays that can display 80×40 column text exist, but are like $10,000…

  12. I know this sounds like a funny (funny haha not funny weird) idea, but what about a matrix of electrical contacts that give you a tangible but not painful current? If fingers are sensitive enough to read mechanical signals, what about electrical signals?

    1. This is a good idea, but the way humans sense pain, electrical shock, etc. might be a hold up.

      Part of the functionality of braille is that it is durable enough to be quickly ‘scanned’ by a person’s finger, the horizontal motion seeming to be part of act of reading – so the motion of the finger itself helps reading resolution, or something like this.

      Something similar to your idea is being looked into in another hackaday prize entry – in the form of a forehead mounted vibratory\tactile array – hooked to a ultrasonic distance sensor.

    2. Pain radiates, I don’t think you could perceive it sharply enough. And even a tingle is unpleasant enough to spoil your concentration. Save it for the kids who misbehave in class.

  13. 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.

    1. I think because of the vast, among blind people, existing knowledge of Braille. And the idea here is to increase the size of the screen, to allow users to scan a page. Reducing it down to one character is the opposite. A one-character Braille cell, equivalent to your idea, probably wouldn’t be much more expensive than your idea.

      Braille typewriters do use something like chording though, they look a lot like chording keyboards, obviously each main key corresponds to one of the six (sometimes eight) dots in each character. You can similarly get Braille computer keyboards where the alphabetic area is replaced with these chording-type keys. The other keys, CTRL, F keys etc just have Braille labels on them.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.