Best component for project conceived

[GaaraIchimaru]

Hi all,

I'm a university student currently reading computer science. I've had this idea for a project, it's essentially an array of motors which need to move up and down (it will be programmed so that some of the motors move less or more than others) creating a 3D surface effect.



Here you can see an example, each of the grid squares would be an individual 'pin' which would (I assume) need to be driven by its own motor.

Although I have had a lot of programming experience, I have near to no robotics/electronics experience, so I was wondering if I could get some feedback and suggestions.

What would the best component be to use? Budget is an issue and if I used the cheapest motors I have seen (around £1.30 each) I would be looking at spening around £130 for a very very small prototype. I don't know if pistons or something are feasible; any ideas?

Thanks for taking the time to read through, feedback, as I say, would be greatly appriciated.

[bdeuell]

something like ... this http://www.gizmag.com/mit-3d-remote-display/29766/

A few details might help identify a good solution for your problem:

Do you need position control or is the state of up or down sufficient?

If you need position control how accurate do you need to be?

What speed do you want the pins to operate at?

How much force do you want the pins to be able to apply?

What pitch to you want between the pins?

[GaaraIchimaru]

Yes, that link provided some ideas, but I'm looking for a slightly more accurate representation i.e. I need the pins to be much smaller, like one of these http://www.tinydeal.com/Big-Size-3D-Image-Sculpture-Creator-Pin-Point-Needle-Impression-Intelligent-Toy-Art-Frame-Gadget-FTY-23585-p-19698.html?currency=GBP&gclid=CJW4g5qFmrsCFWnjwgod6HgArg

I do need position control, I was thinking that I could manipulate pins individually, using motors to turn a certain number of turns to keep the pins up various degrees.

The pins need to reach their position in about 0.2 of a second as to simulate real time to the reciever.

The pins should be able to withstand a force of about 20 Newtons (2Kg).

Hope this information helps Thank you

[Gertlex]

To be blunt... Either loose the force requirement (and probably the speed requirement), or get a different project idea.

The most compact solution I can think of is a ton of weak solenoids.  These could be self-wound, I think, but I've never done such.  They would still be a wiring nightmare that I couldn't tackle off-hand.

[bdeuell]

Here are some thoughts...

Do you really need each pin to deliver 20 N of force?
The toy you showed looks like it has a pin density of about 100 pins per square inch. If you could achieve a pitch half of what they have you would have a pin density of 25 pins per square inch. At 20 N (4.5 lbf) per pin that would apply approximately 775 kPa (112 psi). 


I like the idea of having pins that could screw in and out as it would provide some gear reduction. If you had 1" long pins with a #4-40 thread you would need a motor to turn at 12000 rpm to achieve a 0.2 s response time. Perhaps some of those little vibrating motors they use in cell phones would fit the bill.

Precise of the positioning of the pins will be challenging without a feedback mechanism. The pins could be zeroed to start but error will accumulate over time.

The pins could be hex or square in profile to keep them from rotating with a thread inside for a screw to move them in and out. Hex standoffs come to mind for a prototype anyways. I ran across this perforated sheet with hexagon holes http://www.mcmaster.com/#92725t42/=pokjde density may not be quite as high as you want tho.


Another idea is to use an electromagnet and return springs on each pin. The position could be controlled with the current. However the forces will likely be much lower than desired.

[hobbes]

if each one of the pins is inside a tube you could use forced air or liquid to push them up. Just have one pump pumping and use electromagnets to stop/freeze the pins once they get to their desired location. The electromagnets would act as brakes pushing against the side of a pin to stop it. Very small speakers might work as these electromagnets. A cheap way to monitor the pin positions would be with two cameras, using computer vision to monitor their location in space. Don't forget the pins are two sided so you can hide the cameras in the back and if you color code the pins that are next to each other the cameras will have an easier time picking the same pin. Or you might just get away with timing but the air flow may not be regular enough. A cheap way to control the electromagnets would be using a grid of light sensing diodes ( mini photocells ) placed on top of an lcd monitor. You can than output using DirectX or OpenGL to the monitor, once a pin gets to its location draw a rectangle on the lcd where the light sensing diode is to stop the pin. On the next frame just stop the pump ( or have a valve diverting the flow ) and release the electromagnets and let gravity return the pins to their places, you might need springs if the pins are not facing up or the gravity is not strong enough to return them in time.

For the cameras you could use the ps3 eye they are very cheap now I saw some on clearance for 11$ recently. They have very high frame rate, but they are black and white so instead of color coding you may need to use shades of gray.

here is a link to ps3 eye being used for 3d stereo vision:
http://hangyinuml.wordpress.com/2012/07/19/stereo-vision-grab-frames-with-opencv-2-x-ps3-eye/

[GaaraIchimaru]

I like this idea a lot, air tends to be cheap haha. So the pump could be controlled electronically? Can small enough pumps be bought?

Would the electromagnets stop one pin each or would one electromagnet freeze all the pins on its own?

Say I program the pumps to force the pins up a certain amount each then could the electromagnet be switched on to freeze the pins and then the pumps be turned off and the pins stay in place?

In which case no camera would be required.

Thanks veryone for all the feedback, I'm starting to get a good idea of how I'm going to achieve it but it's still a massively complicated project.

[jwatte]

I believe you want one (large) pump and lots of small valves, not lots of small pumps.
Valves need solenoids. They still cost money.

In general, I think your idea and your budget may not be entirely compatible.

[GaaraIchimaru]

I believe you want one (large) pump and lots of small valves, not lots of small pumps.
Valves need solenoids. They still cost money.

In general, I think your idea and your budget may not be entirely compatible.

How would I electronically allow more air through certain valves and where would the solenoids come in?

[Gertlex]

There's solenoids, and then there's solenoid valves    They're really the same thing, just the latter comes packaged as a valve.

(yes, it's confusing, initially)