Best Actuation Mechanism for Palm-Sized Dynamic Shape Display With 1000+ Linear Actuators

I am looking to build a personal-portable "dynamic shape display". It actually comprises of a grid of hundreds of linear actuators stacked together to make 3D shapes. My main issue is size constraint. It should not be any larger than a laptop. Ideal size would be equivalent to a palm. The number pf actuators is at least 1000.I need ideas and suggestions on which actuation mechanism should i use. 

Muhammed - from your description it seems you're trying to make an electromechanical version of those toys you get that you press something (eg your hand) at the back  and move lots of tiny rods which move out the front and create a "copy" of your hand?  Nice idea, but I don't think technology is up to this yet.  the best way I can think of doing it is to have one linear actuator on an X-Y positioner, which would move the actuator to each rod in turn and push it the required amount, then retract and move on to the next rod.  Or maybe have the rods as threaded rods and use a rotating "screwdriver" to engage each rod and screw it the required distance.
I really don't think you would find linear actuators small enough to give you the required density you're after.  Can anyone else prove me wrong?

you are right. Thats what im 'kinda' trying to make. There is something called "squiggle motor". It is tiny but unavailable in the market and has a rather complex control mechanism.  I was looking for some smart, quick, DIY method that i could use to make my own actuators. Something like a surface that can be morphed elctromagnetically or a huge bunch of tiny solenoids packed together.....

Muhammed - I can see where you're coming from (and where you want to go)  but I think you will struggle to get there.  Your rods will be very close together and there won't be much space for the actuator mechanism.  You remember the old print heads on dot-matrix printers?  And how big they were just to control a few tiny rods that actually did the printing?  But I hope you can prove me wrong - you'll make a fortune!

Hi Muhammad,
I worked on a previous project where we needed to get a really small actuator into a retrofit product. We weren't able to find anything quite small enough. In the end we used Flexinol (http://www.dynalloy.com/flexinol.php), it's a type of muscle wire. It's an alloy that when current is passed through it, it will shrink in length. This can be used as an actuator. Another muscle wire is Nitinol, which will return to it's original shape when current is passed through it.The only problems with it is the current to activate it and the operating temperature, which is likely to melt any nearby plastic. But if you use high temperature plastics and work out the operation, you could build something very small.The actuation motion uses a spring to force it out, once the wire is activated, it will pull it inwards. -Mike

Hi Muhammad-When you say "Personal-portable"- how big do you envision this device being? Referring to David's response about the dot matrix impact printhead- is it OK if the mechanism is much larger than the actual "output" of the device?-Rick 

No,  its not OK if the mechanism is large. The whole thing should not be bigger than a laptop.  Ideal scenario would be that its palm sized (that kind of way too much wishful thinking).

Sounds awesome and quite promising. I looked both of these up and Felxinol is waaaaaay too expensive for me. Nitinol  (https://www.alibaba.com/product-detail/nitinol-shape-memory-alloy-wire-use_60607850891.html?spm=a2700.7724838.2017115.50.7702232byRkR1R) is much more affordable. Can you share some details on how to use it, or control it with a microcontroller?

Since in my application, the device has to be portable and usable like a laptop, I dont want the mechanism to be drawing a super huge amount of current. Would it be feasible for my needs?

Hi Muhammad,
They are both SMA (Shape Memory Alloy) and they are formed in different ways. The Nitinol would need to be formed in the shape you want at a temperature as high as 300 degrees Celsius. When cooled, you can bend it any you want, but once you apply a little temperature (20 to 40 degress Celsius) or current through the wire, then it will return to its original form.Flexinol is already pre-formed. Upon adding heat or passing current through it, it will shrink in length from 6-8%. The passing of current through the wire is what heats the wire, they are of low resistance (typically 6 to 10ohms). At 5V this would mean 0.5A up to 0.833A. This is a little beafy alright, but you can pulse this or PWM this wire. Control is through either a high side or low side transistor (BJT or FET). In situations like this, I typically like low side switching P-Type FETs, but you have to beware of the RdsOn value. If it's too high, it will become a voltage divider and more power will be drawn by your FET than the muscle wire. To be honest, I'm not really sure if it is suitable for your project. It might do the job, but there's a lot of problems to sort out with it.-Mike

The problem is that you have to move something to change the shape. I don't know of any way to move actuators without using significant current. Now you could reduce the power needed if you could latch the actuators in position after moving them. That way average current could be low. Sadly, i don't know how to do this.Hmm, I just had a thought, maybe use air pressure. Miniature balloons that you add or remove air might work if you could make it small enough...This is certainly an interesting application.

I was looking into this a while back, students at MIT have made a few things but nothing in the size you are thinking. Project was called inForm, plenty of info if you google it.You will probably really like this guys research, he's using vibration to drive threaded pins which gets rid of most of the motors:https://www.youtube.com/watch?v=0b0cVgzf9As

How fast does it have to be? Could you use an array of grub screws with something like a CNC machine underneath them? Instead of a single tool, you could use an array of motors driving the pins up and down, but your actuators will be far less dense than your pins. Say you were only able to hit every 20th pin due to size restrictions, you'd have to then retract and step the motors along to  each of the 20 x 20 positions, so it would take a minute to change them all.

I have to change the shapes in real time. I dont really now if this vibration method will be fast enough.

Ah, then the technology you need is 10, or maybe 20 years away, https://en.wikipedia.org/wiki/Carbon_nanotube_actuators

Coincidentally, the reason I am on these forums is I was looking for some help designing electronic muscles, very similar principle to the carbon nanotube actuators I linked, but using a combination of rare earth and electromagnets, kind of like the guts of an electric motor reconfigured to stretch instead of spin. Unfortunately, even if they work, they wont scale down as far as you need them. Magnetic forces scale inversely to distance squared, which is great, but it seems the force you can actually generate scales by volume (aka distance cubed) so they'd be very weak at tiny scales.