Explore Interactive Art with Servomotors & Wood

[dhruv.tara]

I was browsing through you-tube and found this video..

http://www.youtube.com/user/OUlearn#p/c/E4037752E877E4E5/11/BZysu9QcceM

Its seems cool to me, one thing that is bugging me is how many servomotors did this guy used?

This video has sequential vids, I think 3 more... this guy talks about having many servomotors but not the exact number... Are number of servomotors equal to number of pixels?

In that case wouldn't we end up using too many servomotors and hence a good amount of cost?

Plus does anyone has any idea on what kind of wood these guys used? Something really this polished?

 

[schip666!]

I presume one motor per wooden pixel, which would probably make the builder a trust fund baby. More to the point, how many micro-controllers at -- maybe -- 24-32 PWM channels per controller were used?

 

[dhruv.tara]

I don't know about microcontrollers yet... Will be introduced to them formally in upcoming semester...

But I was wondering if we could there control a row of pixels via just one motor... maybe have some controlled gear train mechanism system then... If we discretized the rotational values of wooden pixels?

Say it roates either -45 ... 45 in steps of 15 degrees, then can we use lesser amount of motors?

 

[KingNothing]

Sure, you could do that. Technically (not practically) speaking, you could have a single motor - so long as the motor can resolve as many discrete states of angular rotation at its shaft as there are discrete states of the collection of "pixels".

I am guessing that it is done with each motor controlling multiple pixels, but who knows. One could mass-produce servomotors each with their own custom circuitry, and use some sort of coded signal to talk to each motor individually. I don't think the signalling would be much of a problem.

 

[dhruv.tara]

So what do you guys think can be the total budget of the project? I guess the motors are the significant cost factor or could there be anything else that weighs more than the cost of these motors?

 

[KingNothing]

Keep in mind this is all speculation...but probably $2-3 thousand. The cost is probably mostly in the mechanical movement system. Not just motors, but the mechanisms attached to the motors.

 

[schip666!]

But I was wondering if we could there control a row of pixels via just one motor... maybe have some controlled gear train mechanism system then... If we discretized the rotational values of wooden pixels?

You would either need some-un-specified-thing to engage each pixel to the motor shaft at the right time...maybe a solenoid, or -- as KingNothing suggested -- find some way to encode all the positional combinations of the pixels involved.

When this thing first came out I estimated (actually I forget how many, but let's say a minimal version) 25 by 25 == 625 pixels with one hobby-servo-motor each. I just saw used servos at Goldmine Electronics for $5, so about $3125 for motors (I seem to remember originally coming up with $10K for motors, so maybe it was actually in the region of 40x50 pixels). Then controllers, I'd guess one micro (like the Arduino) for each 16 motors -- so about 40 boards at about $25 each == $1000. Then a power-supply, all the mechanical gim-crackery and materials, plus your free labor... I'm sure you could save at least half the controller budget by rolling your own, or maybe coming up with some kind of timesharing scheme.

And while it's a cool-looking-thing, IMHO, it's not really advancing the art of technology very much. I much prefer machines who have some control over their own expression.

 

[dhruv.tara]

Well I haven't really established anything practical yet (never did an experiment outside the scope of the lab manual) and seeing it for the first time I thought it would be fun. The cost thing bugged me and now I think kind of prevents me to try it...

But actually this summers I shall be doing my training at NTPC (National Thermal Power Corporation, India) Maybe I get get them finance this project for me... :D

 

[schip666!]

I'm sure you can get your summer internship provider to fund the project...haha...

How about starting with something a bit simpler? Maybe a little robot car or walker? I think, with some trickery, one could use 4 of the servo motors to generate a walking motion. And then you could add an accelerometer and get it to teach itself to walk...

 

[dhruv.tara]

Maybe... Can you tell me more about the accelerometer thing?

My friends had made a LFR (line follower) the previous semester in our college fest... I was busy in other stuff so didn't take much part in it... I just had an overview of what they were doing...

Actually graphics interest me a lot so I have learned a few software graphics package... I also love programming and again graphic programming... though I am almost a novice, I would like my project to be somehow related to it then...

That's why I was looking for the video I mentioned... Seemed like I could get to know some image processing and how he used those gray scale pixel values to drive the motors...

By the way can anyone here argue that what are the maximum amount of colors this particular scheme holds? Can we increase them with some simple method?