>>Absolute encoder gives you position, typically in something like
>>a Gray code (a binary code which, like the usual, has 2^N states
>>for N bits, but only one bit changes at a time, which makes
>>things a lot easier- they can easily be converted back and
>>forth).
>>
>>does this mean there is a unique number/output for each spot
>>on the encoder?
>
>Yes, otherwise it cannot give you an absolute position.
>

The methods can vary between encoder types. Here at work we use an
absolute encoder that connects just like an incremental one. The=20
signals are A,B,Z, and their complements. During operation, you get=20
quadrature pulses on A and B, 90 degrees out of phase, Z gives a=20
single ref pulse per revolution.
The "absolute" part comes on every reset. The encoder then bit bangs
a serial port on the A line, giving total revolutions from ref point,
then position within current revolution, all as ASCII

The result is an encoder that has the lean wiring of a relative unit,
but with the advantage of not needing to home on power up.

The units I describe are on five of our robots, and are probably
proprietary, but the info may get you thinking. It would be fairly
simple to implement one of these with a small PIC and the internal
EEPROM..

Lyle

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.