Instructables

N00tron 3D Spherical Volumetric Display - Interactive Exhibit

Featured
n00tron is a unique technology and art exhibit. Generative light art is created in the spherical volumetric display with complete interactivity.  Participants have realtime control of the display drawing algorithms using knobs and switches on the control panel. Great fun for all ages, learn about Persistence Of Vision display technology while producing your own animated artistic creations!

n00tron needs a home in a museum near you!
Please contact us if you can help: dan at MonkeyLectric dot com

by MonkeyLectric

As seen at:  Maker Faire 2010, SigGraph 2010, Maker Faire 2011, Maker Faire NYC 2011


 
Remove these adsRemove these ads by Signing Up

Step 1: What is it?

About n00tron:

- it is a true 3-dimensional display (also known as a Volumetric display)
- the display volume is spherical, about 550mm in diameter
- the display is created by spinning a bicycle wheel with rows of lights inside it
- the bicycle wheel is spun in 2 axes simultaneously
- n00tron is great for making light art, but it is not possible to make 3D bitmaps.
- n00tron is controlled interactively from a control panel
- n00tron draws colorful patterns using simple parameterized drawing algorithms (also known as Generative Art)
- the control panel has lots of knobs that set the basic parameters of the drawing algorithms
- participants of any age can easily create their own amazing 3D light sculptures!

Read on for more information!

Step 2: The Wheel

We used a standard 26" bike wheel.  Its a beefy model with a solid axle and a 6-bolt disc-brake mount.  We bolted a 22 tooth bike gear onto the disc-brake mount.

Mounted inside the wheel is our MonkeyLectric Video Pro display hardware.  We're using standard Video Pro hardware with custom firmware to generate the art.  The display hardware has 256 full color LEDs arranged onto 4 sticks.

Step 3: The fork

The fork holds the wheel just like the front wheel on your bike.

We welded a heavy-duty custom fork to hold the wheel.  The fork is perfectly straight (no rake like a normal bike fork).  A motor mount is welded in.  The fork tubes are cro-moly with 2mm wall thickness.  Total fork length is about 400mm to the headset, plus a 350mm head tube.  The motor plate and axle mount plates were cut with a cnc waterjet from 4.5mm steel.

Step 4: 1st axis drive

The 1st rotating axis is the axle built into the bike wheel.

The 1st axis drive motor is mounted to a plate in the center of the fork.  The mount slots allow us to shift the motor a bit to get the chain tensioned as well as balance the fork weight.

After the wheel, chain and motor are attached and ready-to-run, we balance out the axis further by bolting small weights to the motor plate.

The 1st axis uses a standard bike chain.

Step 5: The 2nd axis

The 2nd rotating axis ("outer axis") is a standard bike headset bear.  Its just like you are spinning the handlebars on a bike really fast, causing the fork and front wheel to spin around and around.

Instead of a full bike frame attached to the fork, there is just a simple tube.  It's like the head-tube on your bike but without the rest of the bike.  We started with a 12" long tube of 6061 aluminum, and used a lathe to machine the headset bearing grooves.  We put in the beefiest bike headset we could find.  The aluminum head-tube is 2" diameter with 3/8" thick walls.

The 2nd axis motor is mounted on slots similar to the 1st, allowing chain tension adjustment. Stock drive gears and chain connect the motor to the end of the fork tube.



Step 6: Main mount plate

The main mount plate holds:
  1. the head tube bearing assembly
  2. the base mounting tube
  3. the 2nd axis motor
The head tube and the base tube are bolted onto the plate using heavy-duty pipe hangar clamps (normally used to hold large pipes to the wall of a building).

the main plate is 6mm thick 6061 aluminum.  it is a 2D cutout shape using a CNC waterjet cutter.  size is about 300mm x 400mm


Step 7: Motors

The 2 motors are pulled from an old wheelchair.  They are 24V DC, 5A motors with a built-in 10:1 gear reduction.  Output speed is about 300RPM, which is very close to what we wanted on our display axes.

Step 8: Slip ring

A slip ring at the end of the fork tube gets power into the rotating fork to power the 1st axis motor.

The LED display itself has lithium-ion batteries mounted on the wheel, in the future we may add a 2nd slip ring so everything is powered from outside.

Step 9: Base Pyramid

The base needs to withstand extremely large vibration forces from the rotating display.  We built the base using standard industrial handrail fittings ("speedrail").  These are very convenient to set up and break down, using just set screws in the fittings.  We can put up the base in under an hour, and the whole n00tron in about 90 minutes.

Most of the tubing is schedule-40, "1.5" guage size, meaning the OD is 1.9" and the walls are about 0.1" thick 6061 aluminum.  The main central tube is thicker schedule-80.  The pyramid base is 5 feet on a side, putting the bottom of the main mount plate about 5 feet off the ground.

Under the 4 corners are solid rubber blocks, and we add about 100 lbs of sand bags around the base of the frame.  The pyramid is very rigid, but the forces at the mount plate are extreme and capable of visibly bending the central tube.  Hitting the right motor speeds can make the entire exhibit walk across the floor due to the vibration!


Step 10: Control panel

The control panel is built into an aluminum briefcase.  very sleek!  We laser cut a custom control panel plate from black acrylic.  Inside the case are 3 Arduino's, each one connected to 6 analog knobs and digital switches.

We bought most of the knobs, the switches were scavenged from old government hardware.

The 3 arduino's connect by USB to a Macbook running Processing.  A simple Processing script reads the knob and switch values and sends them to a Zigbee radio on the serial port.

The spinning display hardware has a matching Zigbee receiver on board.

Step 11: Power supply

The power supply is a dual 30V variable DC supply.  10A capacity on each channel.  We found one with all-LED display so we could see it in the dark room with the exhibit!  We normally set the motor speeds to be constant, about 200 RPM on the outer axis and 250RPM on the inner axis.  The variable supply lets us stop and start the axes now and then to show everyone the hardware.

Step 12: Its done!

Installation photos from Siggraph 2010 and Maker Faire 2010

Step 13: Do your Homework - Gyroscopic forces.

Mechanically speaking, the main question about n00tron is:  will it fly apart into a million pieces killing everyone around it?  our mechanical engineer calculated the forces on the axles and bearings at different rotation speeds to answer this question.  With a dual-axis configuration like we are using, very large gyroscopic forces are produced no matter how balanced each of the individual axes is.  This is a fundamental limitation of any 2-axis design.  The inner axis produces hundreds of pounds of force at the axle, the outer axis is somewhat less.

The gyroscopic forces rotate along with the axes, this results in considerable vibration of the entire structure.  This is the reason for the overall beefyness of everything and the large pyramid base.

Calculations are attached.


Step 14: Can n00tron display princess leia floating in the air?

no.  to display a 3D bitmap image the display mechanism needs to pass through every point in the spherical space at least 10 times per second.  while this display is 3D, at the current rotation rates it only hits about 5% of the points in the 3D space on each rotation.  It is great for making generative art but not good for bitmaps.  To make bitmaps the axes would need to rotate 10x or 20x faster than they do now.  From the last step we know that such a fast rotation would cause the display to fly apart into a million pieces.   There are plenty of ways to make a floating 3D princess leia, this just isn't one of them.

Step 15: N00tron first test

video of our first spin-up of n00tron.  using a simple test pattern we slowly start and stop each axis separately, then do both together so you can see the results in detail.