The Bidwell effect is a remarkable phenomenon of colour vision. The apparatus consists of a disc, one half of which is white, and the other half of which is black. At the junction, a cake-slice is cut out of the disc. The disc is lit strongly from the front with a white light, and rotates a few times a second. A red light is placed behind the disc so that it is revealed every time the gap in the disc passes in front of it. Astonishingly, the red light appears greenish-blue. Not a trace of redness is left.
I first saw this effect as a lecture demonstration when I was an undergraduate. I built this apparatus to see if I could reproduce the effect. If it looks like a pile of junk, that's because most of it is. But it works, and could be the basis of a design for a real-life exhibit.
The important variables are the intensity of the white light on the front of the disc, and the speed of rotation (about 3 revs/sec). Also crucial is the direction of rotation: the user must be presented with the sequence black-red-white-black-red-white. In a real-life exhibit, I would allow the user to start and stop the disc, but I would keep the intensity of the white light and the speed of the disc fixed. It's important that they be allowed to stop the disc, so that they can see that there is no filter in the gap in the disc.
Details of construction
The disc was made of two CDs. Each had a cut-out in it, and by adjusting the overlap of these cut-outs, I could adjust the size of the resulting gap. Half of the front CD was covered in black velvet, and the other half was painted matt white. The front of the disc was lit by a desk lamp, and I used a low-wattage red light bulb as the red light. This was placed behind the disc in such a place that it was hidden except when the cut-out passed in front of it.
The disc was rotated by a stepper motor salvaged from an old ink-jet printer. I used a stepper motor because it gave me very good control of the speed of the disc. It also allowed me to control exactly where the disc stopped. This was useful because I could ensure that the disc always stopped with the cut-out in front of the red light, allowing the user to easily see that the light really was red.
I generated control signals for the stepper motor using a PIC microcontroller - these signals were inputs to a Darlington array (ULN2003 chip) that switched the actual currents to the stepper motor. The program for the PIC chip had a number of features. Firstly, it allowed the user to adjust the speed of the motor. Secondly, on starting, it slowly ramped up the disc speed, because the stepper motor will 'slip' if it tries to go to full speed immediately. Finally, when the user presses the 'stop' button, the program ensures that the disc comes to rest with the cut-out in front of the red light. To do this, I used an infra-red LED/diode combination (OPB625) to detect the cut-out on each revolution. When the 'stop' button is pressed, the program waits until the cut-out passes the detector, and then starts a slowing-down sequence that lets the motor do a predetermined number of steps before coming to a halt. The number of steps, and hence the stopping position, is adjustable using a potentiometer.
Why does the effect happen?
As far as I am aware, nobody is completely sure. However, it is something to do with the after-images that you see after viewing a coloured light. After a red light is extinguished, you see a very brief positive (red) after-image followed by a longer-lived negative (blue-green) after-image. Somehow, the bright white part of the disc that immediately follows the exposure of the red light drowns out the positive after-image. But by the time the negative after-image occurs, the disc has turned so that you have a nice dark background to see it against. What I don't understand is why the initial (real) flash of red is not seen.