Around 1926, John Logie Baird was able to transmit moving pictures over wireless, using a flashing light and a Nipkow disk to reconstruct the image.  A Nipkow disk is simply a spinning disk with pinholes punched in it in an inward spiralling pattern.  As the disk spins, the light flashes in sync with the scanning holes, producing a flying spot of varying intensity that can be used to raster an image by taking advantage of persistence-of-vision effects.

I decided to attempt to produce an updated version of this system using the facilities in the Fab Lab for my final project.  Additionally, I attempted to make it have the capability of projecting its image onto a wall using a laser or bright LED.  I was met with limited success, but some interesting observations were made, and I have confidence that I will be able to get the system to work.

I designed the main component of the televisor, the spinning disk, in Solidworks.  It has 32 pinholes, enough for a 32 vertical line display, as well as 32 larger holes to server as an optical encoder.  I made the frame for the televisor out of laser cut acrylic, and designed a control board, as well as phototransistor and IR LED daughter boards in Eagle.  I attached these to the frame on either side of the spinning disk so that I would get a nice square wave as the frequency of the disk changed.

The control board includes an ATTINY44 microcontroller to control the imaging element and IR LED as well as read input from the phototransistor.  It also uses a MOSFET to drive the disk’s motor.  I spent quite a while burning out MOSFETS before I realized that a diode was needed to be placed across the terminals of the motor in order to shunt the nasty back EMF from the motor back up to the power rail when it appeared.  I also added a much larger MOSFET to the design.

The final revision of the circuit board incorporates this larger MOSFET as well as a stereo phono jack for getting analog signals into the microcontroller, as the NBTV club in England distributed .wav files containing sync pulses and content for televisors such as the one that I have constructed.

The microcontroller reads the time between pulses from the optical encoder and adjusts the duty cycle of the motor PWM up or down in order to control the rotational speed of the disk to 750 RPM.  The system oscillates a little bit, but it is within an acceptable range.  The real problem (and the problem that kept me from finishing this project on time) was that the microcontroller seems to lock up after an indeterminate amount of time, and nobody I can find, in person or online, seems to know why.

Additionally, I did get some rudimentary PWM of a beefy LED working, but I messed up all of my aluminum stock for disc-making in the waterjet before I figured out what how to fix my toolpath problem.  I experimented with making discs out of cardboard and acrylic, but the cardboard was too thick and the acrylic was too transparent to get reliable results out of the optical encoder.

However, when holding my diffused LED up behind the disk’s malformed holes, there are 32 distinct regions of repeating patterns around the annulus of the disk.  Of course, this points out the imperfections in the disk’s manufacturing process as an ideal disk would produce a uniform raster brightness, but this does show that spinning holes and a flashing LED can produce an image.

I also experimented briefly with using an uncollimated laser beam to project an image through the disk.  This was unsuccessful due to the poor brightness uniformity across its spot as well as the tendency for the laser diode to overheat and fade out when running it from my bench top power supply.  Making a projector using a bright LED or lamp should be possible, but would require use of optics to focus the image.