Bristol Braille Technology

Technolgies and Design Challenges

append delete Tom L

Hi Ed,

Tom here, I just wanted to open up the topic of what solutions are being used moving forward to solve design issues.

In your demo video you have mentioned locking mechanisms that will be much smaller for use in the final product. Could you go into further detail as to what they are and provide a link if it is an off the shelf product.

Have you considered using electromagnets to hold the arms in position? A solid iron rod spanning across the width of the display with individual coil clusters for each arm might work a treat and hopefully cut costs.

You could remove the teeth from each arms design and replace that with a metal coating. When each arm reached the correct position we could fix the arm with the magnet.

The next question is how you are planning to slide the arms back and forth, what will you use to actuate that movement? Watch this video i might be or interest :)

http://www.cs.cmu.edu/~tmashimo/video/RLPA14mm.mpg

Are you planning to use something like self lubricating bearings so that movement will remain smooth for a long time without more than simple cleaning maintenance?

Has there been any consideration given to how to track the extent of that movement?

Anyways that my brain dump for today.

Tom

Reply RSS

Replies

append delete #1. Ed Rogers

Good to hear from you Tom.

Currently we're using open frame solenoids from Geeplus Europe (http://www.geeplus.biz/), however as these are overlarge and overpowerful for what is needed of them we are looking into alternatives.

Yes, we have looked into plain electromagnets as a solution. The problem is twofold; they need to provide enough power to override the magnetic attraction drawing them back, increasing power demand, and they would need to be always on to keep the sliders in place when the machine is turned off.

The advantage of solenoids is that they are essentially the same, except they provide a physical motion perpendicular to the motion of the sliders. This means no power is required to either stop the slider, or keep it there, only to lift the armature up on release.

In time we may adopt something more similar to your idea; i.e. a flat array of electromagnets drawing armatures up and down.

All actuation is done by a dumb DC motor, moving at a relatively steady speed, rotating one way and then t'other. The positions of the sliders are detected by an LED sensor which sets when the next selection of locks is released.

The method used to obtain smooth movement of the sliders is self lubricating plastics being held in place at a minimal number of places.

append delete #2. Tom L

Hi Ed,

Could you explain why the magnets would have to remain on when the unit is switched off?

Is this to do with the sliders moving around during transportation?

What is the problem with them moving?

Might there be a way to lock them using the moving plate with the plns at the front, putting it into a "Park State" when the unit is switched off which lowers onto the sliders holding them in a fixed place?

Minimising the wieght of each sider would possibly make using magnets potentially viable.

I am just throwing out some ideas to generate some discussion.

Cheers,

Tom

append delete #3. Tom L
append delete #4. Tom L

http://www.fipa.com/us/en/Products/End-of-arm-tooling/grippers/magnetic-grippers.php?goto=8052

This might do the Job, I don't seem to be able to find any Electro Permanent Magnet suppliers. Ideally this would be the best route to take if we could find someone to supply them. Otherwise the grippers would do the job quite nicely.

Some further research and talking to the robot developers might point us in the direction of someone who makes those magnets.

Cheers,

Tom

append delete #5. KeithS

Tom,

I have been supporting Ed's open source project, from time to time, with a few suggestions re the overall design.

Most recently we have been discussing how best to document the basic design concepts and this forum is a starting point.

With respect to the latching scheme that allows movement of the slider and then holding it in place, for the presentation of a specific braille character, I would like to reference a preliminary diagram available at:
http://www.kscitech.com/BBT/Slider_diagram1.jpg

We note that 2 adjacent sliders are positioned to present a complete 6-dot braille character.

I should emphasize that this represents the generic design and there are some electro-mechanical details still under evaluation.
In the meantime BBT is proceeding with the fabrication of a "next version" of the Quixote display per this generic design.

We will welcome your comments and further questions.

Keith

append delete #6. Tom L

Hi Keith,

Thanks for showing me the design information. It would be good to pull this all together as I now have a better understanding. I had suggested a wiki as they tend to grow move organically than a forum.

What strikes me as being overly complex in that setup is the braille pin assembly. Can I ask why there is this additional up and down movement and not a smoother ramp used with rounded end pins so they slide up and down. This would alleviate the need for the mechanical up and down movement of the pin board assembly. Also the overall display refresh would be faster without this added movement.

The less moving parts the better, is there a reason why the pin board had to move? This would also reduce cost, as long as the end of the pins could be coated in the same self lubricating materials being used else where in the design.

The only disadvantage is that the force needed to move the slide encoders would be increased.

What i also want to establish is how set in stone this design is and if any of the ideas I am generating could be considered in future generations. Maybe I have come in too late on the project to be adding in ideas.

I am happy to draw this idea if it helps to illustrate what I mean.

Cheers,

Tom

append delete #7. Tom L

Bistable Solenoids

A bit of research it seems the you might be able to get these at the right size and with open cage design you can cut costs. This means they will not last as long but will be cheaper.

http://www.cedrat-technologies.com/en/about-us.html

http://www.alibaba.com/showroom/bistable-solenoid.html

http://www.motran.com/about-us.html

http://www.directindustry.com/industrial-manufacturer/solenoid-78963.html

http://www.movingmagnet.com/en/bistable-actuators-ON-OFF-solenoids/

append delete #8. KeithS

Hi Tom,

You wrote:
> What strikes me as being overly complex in that setup
> is the braille pin assembly. Can I ask why there is
> this additional up and down movement and not a smoother
> ramp used with rounded end pins so they slide up and down.

You have nicely summarized the alternative modes of operation, but let me comment as follows:
- The reason for moving the braille pin assembly relative to the sliders (in the generic diagram it is lifted up and down) is to isolate the sliders from the pins during a refresh cycle.
- We note that there may be an external force acting at the head of some braille pins, such as that associated with a braille reader's finger resting on the display surface.
- The specification for completing a refresh cycle is 500 msec.
Thus it's imperative that any forces acting on the sliders in opposition to the linear actuation force, moving the sliders both forward and backwards, be kept to a minimum. Typical counter-acting forces are, for instance, due to the mass of a slider under acceleration and frictional forces due to contact between slider surfaces and the housing.

With respect to design features being cast in stone, I can not speak for Ed, but I expect that all ideas should be documented and filed away for future reference by BBT. That being said, Ed's top priority right now is to proceed with a 'baseline prototype' based upon the current BBT design.

More later,
Keith

append delete #9. KeithS

Tom,

I have added a label for the 'lifter' feature on the slider-encoder to our reference diagram.

Now I need to add to my previous post.
You commented:
The only disadvantage [smoother ramp used with rounded end pins so they slide up and down] is that the force needed to move the slide encoders would be increased.

You have pointed out a major disadvantage with attempting to force the slider under a pin mounted in a static pin board assembly.
The fundamental problem is that the force (vector) required to lift a pin into the up position is orthogonal to the force moving the slider. So the lifter feature on the slider has to be some form of ramp to implement a translation of forces. However, this is difficult to implement at braille cell dimensions because the distance between adjacent slider states (PS1 through PS10)is a nominal 2.5mm. Thus there is always a significant force trying to jam the pin sideways in the lower through hole in the pin board assembly as it moves in the upward direction.

In the generic design under discussion, however, once the silder has been positioned for a specific half-cell dot pattern the force (vector) moving any pin upwards is colinear with the pin axis - an ideal situation.

I trust that this clarifies things, but I may well have overlooked something - so please feel free to continue on this particular topic.

Keith

#10. KeithS

This post was deleted by its owner

#11. KeithS

This post was deleted by its owner

append delete #12. KeithS

Tom,

You wrote:
> Thanks for showing me the design information. It would be good to
> pull this all together as I now have a better understanding.

Per discussions with Ed for this Open Source project we intend to give an overview of the operational aspects of the BBT RBD design in the Concepts section of this forum.
There will also be some historical notes related to the evolution of the generic design currently under discussion.

In the meantime, here is a diagram that may help with a further understanding of how the slider-encoder scheme works:
http://www.kscitech.com/BBT/Slider_encoder_schemeC.jpg
[I tried to embed this table in the post but the column formatting was messed up - probably because I didn't use tabs correctly :( ]

This is actually a variation on the original 6-dot encoding scheme, dated May 18, 2011, to match the current design's refresh cycle.

We note that in the 'home' state the slider is pushed all the way forward to a stop in the RBD, i.e. to the right in the previously referenced overall mechanism diagram. For this slider state, if the pin board was lowered (to complete the refresh cycle) the left-side, half-cell, braille pattern would be: dotl= down, dot2= down, dot3= up.

In general, however, the slider is pulled backwards (to the left) to set a specific braille pattern. For example, in the mechanism diagram the slider is in state 5 and the braille pattern is: dot1= up, dot2= down, dot3= up.

More later,
Keith

append delete #13. Ed Rogers

Could you explain why the magnets would have to remain on when the unit is switched off?

Is this to do with the sliders moving around during transportation?

Yes. The difference between the design that Keith has illustrated above and the one you propose is that the former stops the magnets, but does not lock them, whereas the latter does both. The locking may not be needed for desktop usage, but would be vital for use in a classroom environment or while out and about.

Ed Rogers added on

Keith, correct me if I'm wrong, but don't you only need LS1 through to LS8 on your diagram?

#14. KeithS

This post was deleted by its owner

append delete #15. KeithS

_Keith, correct me if I'm wrong, but don't you only need LS1 through to LS8 on your diagram?_

Ed, You are correct. There are only 8 different slider states to be latched. The diagram has been revised (Rev.1.2):
http:/www.kscitech.com/BBT/Slider_diagram1.jpg

Thanks,
Keith

KeithS added on

http://www.kscitech.com/BBT/Slider_diagram1.jpg

append delete #16. KeithS

A first animation to demonstrate a slider in action:

http://www.kscitech.com/BBT/Slider_animation_01b.gif

#17. davet

This post was deleted by a moderator

Reply

(Leave this as-is, it’s a trap!)

There is no need to “register”, just enter the same name + password of your choice every time.

Pro tip: Use markup to add links, quotes and more.

Your friendly neighbourhood moderators: Ed Rogers, Steph