Incremental Optical Encoder
Incremental Optical Encoder
Optical encoders can be divided in to two main groups, that is absolute and incremental transducers. Both types transform the mechanical input quantity, namely the physical angular position of its shaft, in to an electrical output quantity by light absorption. This principle is illustrated in Figure 1. The disk can be either transparent, such as clear plastic or glass, with opaque lines printed on it or made of metal with slots cut in it.
Optical encoders are widely used for the measurement of revolute joint positions in robot systems. Their main advantages are high accuracy (a 12 bit optical encoder will have an inherent reading accuracy of 1/4096 = 0.024%) and virtual absence of mechanical wear in operation due to the non-contact nature of the transducer. It should be noted that for each optical encoder there is a roughly equivalent magnetic position transducer, which also does not suffer from mechanical wear, but that the use of optical position transducers is much more widespread within robot systems.
However, this technique is somewhat inflexible because it requires prior knowledge of the total number of lines on the disk, so that the counting hardware can keep track of the number of complete shaft revolutions. Therefore changing the position measurement resolution, that is changing the total number of lines on the disk, would also require a change in the counting hardware.
An alternative, as adopted by most major manufacturers, is to use a third LED-photodiode pair and a separate channel containing a single opaque line (or clear slot if a metal disk is used) which then provides the detection of each complete revolution, as required. The output signal from this third channel is in fact a pulse train whose period is equivalent to one complete shaft revolution; each pulse can be used to ‘zero’ the angular position count as well as providing the signal for a shaft revolution count, as shown in Figure 1 in the section on interfacing.
One drawback of the incremental encoder systems thus far described is their poor angular resolution, Δα. This depends on the number n, of opaque lines on the glass disk (or the number of transparent slots on a metal disk), as shown in eqn (2.3), and the width Wp of the photodiode active area Wp on the disk plane, which in turn depends on the photodiode mounting distance from the disk centre, as illustrated in eqn (2.4) and Figure 3.
Note that, since most photodiodes have a square active area, the required dimension Wp is simply the square root of the active area given in the photo diode data sheet:
For example: to design an optical incremental position transducer with a resolution of ±3° one would require an encoder disk with at least 120 slots, that is one every 3° of arc. To match this resolution the photo diode would need to have an active width Wp no bigger than 0.25 mm2 mounted 10 mm from the disk centre. An example of such a system is shown in Figure 4.
There are two main ways to increase the angular resolution without resulting in a larger disk diameter.
The first alternative relies on using gearboxes to yield n disk revolutions (where n is an integer >1) for every load shaft revolution. Since motors are often used in conjunction with gearboxes this alternative in inexpensive and only requires mounting the encoder disk on the same shaft as the motor instead of the load one. The encoder disk would therefore be rotating at the motor speed and the counting hardware would need to be designed to cope with the higher operating frequency; this is not usually a problem since a typical motor speed of 6000 rev/min would produce a pulse train of only 25.6 kHz at the output of a 256-line optical incremental position transducer, well within present digital technology operating frequencies. Any gear inaccuracies, such as backlash, would however, add to the load position measurement error, thus seriously limiting the overall transducer performance.
Another way to increase the optical resolution of these transducers is to use a second stationary ‘phase plate’ between the rotating disk and the photo diode so as to form Moire fringes on the photo diode surfaces. Figure 5 shows the schematic diagram of a commercially available optical incremental transducer based on such a principle. The angular resolution of this device does not depend on the photodiode active area and its related optics but on the Moire line width which can be adjusted to suit the application. Figure 2.11 shows a typical Moire fringe pattern obtained using another commercially available device based on this principle.