-
Hi folks,
First time visitor and poster here but I've been following the EEVblog on Youtube religiously for the past few months.
Anyway, I've had an idea for a low cost Braille terminal for the blind rolling around the back of my head for a few years now and tonight I've decided to release it to the public domain and the interweb.
If you've ever seen what they have the nerve to charge for a lousy 40 character by one line Braille display nowadays I swear you'll mess your proverbial drawers.
What I have here, in my own crude fashion, is a design that shouldn't realistically cost more to manufacture than any crappy inkjet or impact printer.
Anyway, I figured this is as good a place as any to unveil it and release it to the public domain.
Any thoughts, critiques or comments are more than welcome...
Thanks for listening,
Gary
-
So really? 54 views and not a single comment or question?
Gary
-
Still trying to figure out how it works?
You didn't state it explicitly (it is probably obvious to you), but do the pins start out in the UP position, and then the solenoids pull DOWN selected pins?
Can you really do this fast enough for a practical line of Braille characters and an experienced Braille reader?
How noisy is it? How much power is it going to take? -
The solenoids fire upward, pushing the pins into place, then the steel locking tines slide into place to keep them from slipping up or down until the line is refreshed. Meanwhile the solenoid stepper moves over to the next column.
One item missing from the diagram is a third perforated plate between the solenoid and the bottoms of the pins to keep the pins from dropping out. The locking tines would slide between the bottom and middle plates.
I'm not a CAD expert, having grown up with mechanical pencils and T squares, so my attempts at using a paint program as a CAD designer should be viewed with a bit of leniency
It might be as noisy as your typical impact line printer and should operate about as quickly.
I'm not saying it's the quietest or fastest Braille terminal to come down the pike, but compared to the prevailing designs which are effectively a solenoid for each pixel it's GOT to be a hell of a lot cheaper and at the very least should be considered for developing world applications
A 40 character by 1 line Braille display currently prices at around $4000 a pop! An 80 character by 1 line display prices in at $8 Freaking THOUSAND!!!!
I'd estimate the cost of manufacture of my design in relatively limited mass production runs would still average no more than a few hundred apiece. If it became a Braille standard device, costs could be cut to about a $100 or less.
Look at it this way. Current Braille readers, say an 80 char. display, contain about 480 micro solenoids and associated wiring and logic apiece.
My design cuts the active component list to 3 solenoids and a couple of stepper motors and associated control logic. Add in some perforated plates and 480 simple injection moulded plastic pins and you have a refreshable Braille Line Display 80 characters or as many characters long as you'd like.
No matter how you calculate it, that's GOTTA become a huge savings in the production costs and could effectively bring Braille computing to the sightless of the third world.
I read the other day that less than 10% of sightless adults today are even taught Braille at all and that of the small percentage who actually are employed 90% of them read and write Braille. The main reason for the lack of education is the lack of affordable options. Braille hardcopy and terminals are horribly expensive.
Also I read that India is looking desperately for an economical Braille terminal design to further improve the employability of the sightless.
People go blind at all ages from all causes ranging from disease to malnutrition to war and if this can ease the burden of the sightless around the world for less than the cost of a OLPC netbook, then I'm all for it.
Seriously I'm opening this idea to the public domain and make of it what you will, but it really would be huge benefit to the human race as a whole if we could expand the limited options available to the sightless today.
Gary
PS. The reason I'm getting involved in this is that I'm an amateur keyboard player and there are quite a few visually impaired keyboard players out there who would welcome an interface which would allow them to program the complicated beasts which modern synths have become.
I was toying with interfacing an 80 char by 1 line display with a Kurzweil PC3XXX series synth but I almost choked on my coffee when I saw the current pricing for Braille display terminals.
I had this idea years ago, but it never occurred to me to really start pushing it until I started taking a good hard look at what sightless people have to put up with in our modern world.
Anyway, here's hoping that other people online will pick it up and run with it. If there was ever a USEFUL gadget that people really need in our consumer spending driven world, THIS is the one that deserves all our brains and efforts.
-
Hi Gary. First of all, for the initiative
Haven't had a chance to be able to look at the design in detail, but may be able to mock up a 3d view if people are confused how it works to get the point across clearer, but two points for now:
1) Research. Obviously there are a few manufacturers who produce several thousand dollar devices. But have you checked on the initiatives for low cost braille units? No point reinventing the wheel if such a device does exist (a cheap braille device that is) but the issues are elsewhere (e.g. development/marketing/funding etc, instead of the concept). A super quick google indicates there are a few companies developing low cost braille devices (e.g. daisy.org and ukaaf.org). I also found a device called the TacRead. At work atm, so haven't been able to look into these extensively.
2)Production cost. Although a printer would be more complex and is cheap, remember the market for them is huge and as such they can be mass produced cheaply. The engineering that goes into them however is quite large, and as such a small number of units may still be very expensive due to needing to recoup development costs. Plus you'll get quality variance from say a multinational company with hundreds of engineers designing a printer vs a few people with less professional experience making something.
I'll have a review of the technical details this evening and get back to you regards your specific concept.
link: http://www.ukaaf.org/braille/agreement-reached-to-produce-a-low-cost-refreshable-braille-display -
hlokk >>> Thank you very much for the interest and the efforts you're putting into this.
My background is more along the lines of electrotechnology dropout who's been out of the technology loop for decades so some professional input is extremely welcomed.
I've also never had much of a stomach for red tape or the piranha infested waters of the patent offices so it really struck me that my best option for seeing this concept developed was a diasporic approach which might even inspire others out there to make improvements or even come up with better designs of their own.
My own google searches came up with some other approaches, (including some real Rube Goldbergs) but I think my idea has the advantage of an industry proven dot matrix technology combined with a novel approach to fabricating the display using low cost plastic "dumb" components instead of "smart" ones.
One item which I omitted from the drawing which, again, while obvious to myself, might not be so obvious to others is that the spring steel retention tines would be designed to be extremely thin and flexible and thus be able to wrap around beneath the mechanism instead of extending a foot out to the left as my drawing might seem to indicate. It would add thickness to the unit but not excessive bulk.
Please feel free to post any further questions you might require in order to cobble up a more professional drawing.
Regards,
Gary
-
Did a quick sketch of how the thing is supposed to work, but not super clear how the locking pins engage. Can you explain or draw this part up. Think I understand the rest of the concept.
-
It's getting ridiculously late here so I'll draw up a sketch by hand and scan it in tomorrow afternoon or evening.
Thanks again for all your help.
Gary
-
Bad case of insomnia tonight so I put my time to good use.
If I could I'd make this a multiline display but the mushrooming complexity of two locking coils per line and associated stepper hardware render that idea impractical so far.
Anyway, I think I've managed to describe the algorithmic process clearly enough now that anyone could follow the logic of it.
I'd love to hear your thoughts and if you can come up with any other suggestions I'd love to hear them.
BTW, I googled the few open source Braille display projects out there as well as all the commercial competition and I've seen nothing else that even comes close to this idea.
I also learned that India is home to 3/4 of the world's 160 Million blind.
People REALLY need this hardware. If someone out in China or India can develop this and put
it on the market for under $100 they'd be national heroes.
Gary
PS. I've been rethinking the Kurzweil PC3K synthesizer display problem and may have come up with an off the shelf solution. Incorporating a digital buffer piggybacked on the display and using a simple text to speech algorithm to read the contents of the screen to the blind musician thru an earpiece.
I may end up making that my comeback project this winter after almost thirty years out of the electronics game.
But if anyone else out there can come up with the hack any quicker, feel free to drop me a line. I know a couple of blind musicians who are dying to lay their hands on that hack
Gary
-
That is more complex than my understanding of your first diagram and description.
I had expected that you were using a mechanical "latching" mechanism to hold the designated pins up, and then one single "reset" solenoid to clear all 480 pins.
Or even a manual mechanical "reset" lever or button, etc. operated by the user. And activating the "reset" button would trigger the computer to "display" the next line, etc.
-
Well to be fair, I was a little fuzzy with the initial explanation of the operation of the device.
Then again, I think now I've managed to clarify the key points in the design and sum up its operation in a nutshell.
Actually it only occurred to me last night to use a coiled spring ribbon instead of a simple flat strip folded underneath. With the sort of latching mechanism you suggest it would be free to retract under its own stored power when a refresh was initiated. Right now I could go either way with the design.
At any rate, what I'm offering here is a low tech solution to the problem. A simple mechanism based on proven technology using cheap, "dumb" components instead of an expensive array of 480 piezo cells which seems to be the current state of the art.
Like I said earlier, I really hope someone from India or China jumps in here and picks up the ball, since those nations have a solid history of adapting low tech at a rock bottom price point.
I'd imagine with what's been posted in this thread so far, someone with access to a 3D printer would already have enough information to fabricate a working prototype.
At any rate, regardless of whomever develops the end product, if it can be mass produced at the $300 price point or below it would be a huge benefit to humankind as whole.
Gary
-
A sliding plate with 480 tiny leaf springs could "latch" the pins when they were pushed up. Then by sliding the plate, it would release all the pins to drop down (perhaps aided by tiny coil springs?)
The latch-spring plate could be stamped out of a single piece of phosphor-bronze spring material to reduce manufacturing costs.
I think you have a viable idea there. -
How would that work for latching the pins in an OFF state? It would have to be neutral regardless of whether any given pin was either OFF or ON.
Gary
-
Wouldn't the pins be in a normally "off" (down) state unless pushed "up" by the solenoids? (And pushed down by small springs if necessary)
We are assuming it will be used in 1G on the surface of our planet, and not in weightless outer-space. -
Those are the compression springs in the drawing. Sorry but it's hard to draw a helix with a lousy paint program.Quote
A sliding plate with 480 tiny leaf springs could "latch" the pins when they were pushed up. Then by sliding the plate, it would release all the pins to drop down (perhaps aided by tiny coil springs?)
I'm just not getting at what you're driving at with this. I don't understand the purpose of 480 stamped leaf springs on a monolithic plate.
I asked a valid question how it worked and you gave me sarcasm.
Seriously can you do up a diagram of how a single plate works better than two sliding spring steel ribbons?
Gary
-
Incidentally if I think I'm understanding you right, you want to use a single perforated plate to latch all 480 pins simultaneously?
If so it won't work.
The purpose behind the dual out of phase steppers is first a column is set, THEN it's latched, THEN we move on to the next column.
Sliding a single plate into place to latch them after they're ALL set serves no good purpose except to add needless complexity to the design.
And sliding a plate into place after each column won't work once you've passed the first column.
Look at it this way. Each pin has a spring loading it downward to the OFF position.
Each column consists of three pins in a row.
First you fire off the correct sequence of solenoids to set the first column.
Then you latch that column and ONLY that column and THEN you move on to the next column in the sequence. Repeat 160 times for 80 characters.
At that point, ALL pins are latched either ON or OFF and the cycle halts until a refresh or scroll or page button is pressed.
THEN AND ONLY THEN the steel retention ribbons and the solenoid head are retracted back to the beginning of the line to commence hammering out the next line.
A monolithic latching plate with elliptical perforations and or die stamped leaf springs is simply not capable of doing a column by column progression in this fashion. It either locks all 480 pins at once or nothing at all.
The compression springs in both original drawings are almost incidental. It really doesn't matter if they're die stamped leaf springs or tiny coils. The only function they serve is to return all the pins to the OFF position once the line is reset.
Frankly you could use whatever works best for the lowest cost for those simple compression springs. Stamping them into the guide plate would probably necessitate the cost of creating an additional die to create a different plate design from the reading surface and the bottom stop plate. The additional tooling costs again add needless complexity and additional cost to the design for no good reason. Personally I think I'd rather go with simple coil springs and a simple semi automated assembly jig for smaller production runs.
Gary
-
One further refinement that's not in either of the drawings is the fact that the tip of the retaining leaf springs should be bidirectional wedge shaped, ie. diamond shaped, in order to better guide the ribbons thru the locking tabs without getting stuck or hung up somewhere inside the mechanism.
Gary
-
The springs are like a "barb" once you push a pin UP, the barb holds it UP until you release it by withdrawing the barb.
All 480 barb-springs are attached to a single plate.
The pins all start out down.
Then the solenoids go along and push UP the appropriate pins.
Each barb holds its pin UP after the solenoid pushes it.
As the solenoids push up all the appropriate pins, the barbs hold them up until reset.
Then the user has as much time as they need to scan the line, and then initiate the reset.
Then, by moving the plate to the side, the barbs release all the UP pins, to return DOWN, and we are ready to do another line.
All the 480 barb/springs could be stamped out of a single piece of spring metal -
Again, this scheme would only work after the entire line has been written.
If you are relying on sheer friction to "grip" each pin on an upstroke this would necessitate a more powerful solenoid to overcome the friction and the compression spring's resistance combined, plus it would prematurely wear the pin surface or the barb tips with repeated cycles.
Realistically we want something like this to function reliably for at least a million duty cycles and incorporating a friction lock design would induce wear to the point of failure in less than 1/10th of that timespan.
Add to that the additional tooling costs of fabricating a special die to manufacture the gripper plate and the shortened lifespan of the end product and I think my original concept wins out as the better design of the two.
In fact, one of my biggest worries is the long term effects of friction between the ribbons and pins from constant cycling and even I'm not certain about which materials would work best and whether or not lubrication would be necessary.
If we use plastic pins I could see the pins being sawn in half after a few tens of thousands of cycles. In the final design I'd think cast or forged steel pins would work better, even if they are a few dollars more expensive.
Plastic would probably work just fine for a proof of concept 3D printer prototype, but for final production release versions metal on metal would probably come closer to achieving the kind of reliability we've come to expect from dot matrix technology.
Gary
-
The springs are like a "barb" once you push a pin UP, the barb holds it UP until you release it by withdrawing the barb.
All 480 barb-springs are attached to a single plate.
The pins all start out down.
Then the solenoids go along and push UP the appropriate pins.
Each barb holds its pin UP after the solenoid pushes it.
As the solenoids push up all the appropriate pins, the barbs hold them up until reset.
Then the user has as much time as they need to scan the line, and then initiate the reset.
Then, by moving the plate to the side, the barbs release all the UP pins, to return DOWN, and we are ready to do another line.
All the 480 barb/springs could be stamped out of a single piece of spring metal
That's brilliant!
Sent from my Smartphone -
Again, this scheme would only work after the entire line has been written.
It works as each pin is pushed up. Whether you print an entire line or a single pin.QuoteIf you are relying on sheer friction to "grip" each pin on an upstroke this would necessitate a more powerful solenoid to overcome the friction and the compression spring's resistance combined, plus it would prematurely wear the pin surface or the barb tips with repeated cycles.
Right. That is why it is a "latch" If you have ever inserted pins into a connector shell you know what I mean by a "barb spring". Even a door latch is an example. Once the door closes, the latch holds it in place without power or friction until it is released by withdrawing the pin.
Indeed plastic pins would be inappropriate for a final design. -
OK, I'm grasping it now. In order to get a positive "lock" on a pin though, you'd have to machine a tiny groove around the circumference of the pin so it would lock at a uniform height.
Pros : It eliminates the latching ribbons and associated stepper motor eliminating complexity and associated costs.
Cons : Reliability??? MTBF??? It would take an experienced process and materials engineer with 6 months of computer time to run the simulations needed to determine which of the two systems is more reliable and less error prone and less likely to fail catastrophically over the projected lifespan of the unit.
Con : Cost of repair if and when the springs start to fail.
Con : Cost of manufacture. Phosphor bronze switches are generally rated in the hundreds of thousands of cycles instead of millions of cycles before failure and is a far more expensive material than spring steel to purchase while spring steel is far more costly in terms of tooling for production dies. The spring ribbon steel/stepper locking mechanism uses off the shelf materials for lower cost and lower tooling costs.
At any rate it would have helped enormously if you'd simply scanned a sketch of what you were describing in the first place but even though I now understand your concept, it's anyone's guess as to which one is the best mix of cheapest cost and highest reliability.
Just for the sake of others following the thread why don't you draw up your version and submit it here in the thread so others can dissect and analyze and offer their own enhancements as well?
Perhaps some process engineer will step up and offer an informed opinion of which version holds more promise?
Anyway, it's late and I still have a dozen chores to finish before bedtime...
Regards
Gary
-
Cons : Reliability??? MTBF??? It would take an experienced process and materials engineer with 6 months of computer time to run the simulations needed to determine which of the two systems is more reliable and less error prone and less likely to fail catastrophically over the projected lifespan of the unit.
You could just make a few of single pin prototypes, then pulse the pin solenoid and reset solenoid alternately even 250ms to do a 1 million activation test in under 6 days.At any rate it would have helped enormously if you'd simply scanned a sketch of what you were describing in the first place but even though I now understand your concept, it's anyone's guess as to which one is the best mix of cheapest cost and highest reliability.
His initial description seemed pretty clear. Anyhow, I've attached a version of it (quality is very poor). The return springs are shown as being formed from punched sheet, as are the holding barbs. The springs could of course be wire springs, but the punched version would be cheaper and much easier to assemble.
-
Dave >>> Your CAD-Fu is very powerful indeed
This may indeed be a workable design.
I'm really hoping someone out there is following this thread and running with it.
Lots of good ideas coming up here.
Alex >>> The Optacon display is really yet another variation on the piezo approach which is the current state of the art.
Hopefully some manufacturer in China or India will pick up on these ideas and develop and market an Industry Standard Terminal that could be adapted to any potential use from OCR to direct ASCII input with the appropriate modular input device.
I really want to thank all you guys for your interest in this topic to date and all of your contributions.
Hopefully it shall remain Open Source or Public Domain in its ultimate incarnation to help keep the costs and MSRP low enough that anyone could afford one.
Gary