The problem is usually due to two physical properties of an optical encoder. First, some optical encoders have ``weak'' phototransistors. This means the phototransistors cannot let much current through the device, even when it is 100% on. As a result, it is often necessary to use a large value pull-up resistor. This, in return, change the response time of the phototransistor. As a compromise, sometimes it is necessary to use a less-than-optimal resistor value, making it impossible for a phototransistor to yield a low voltage when it is fully turned on.
The other limiting factor, surprisingly, is the encoder disc. Industrial encoder discs are machined from thin aluminmum sheets. A thin foil of aluminum is 100% opague to infrared. However, it is far more cost effective to create encoder discs from transparency sheets (for projector use). A stripe pattern can be easily generated by various programs, and the pattern can be printed to a transparency sheet using a laser printer.
Although the color black from a laser printer is dark enough for the human eye (in the visible spectrum), it is not guaranteed to be opague enough in the infrared spectrum. This is because infrared has a longer wavelength than visible light. As a result, a toner density that is sufficient to block the transmission of visible light is not necessarily sufficient to block the transmission of infrared.
As a result, a cheap homemade encoder disc may be somewhat transmissive even at the opague (black) area. This infrared leak through means that the phototransistor is always on a little bit, and it is never completely off. The corresponding output of the phototransistor never reaches the high end of the rail.
Copyright © 2006-02-15 by Tak Auyeung