ENCODER WITH REFERENCE MARKS
This invention relates to encoders.
Encoders are used to measure the movement of one member relative to another, and typically comprise a scale on one member and a readhead on the other. As the readhead passes along the scale, it reacts to periodic marks on the scale in order to produce a periodic output, e.g. a train of pulses. The pulses are counted incrementally by an external counter in order to give an indication of the distance travelled.
It is known to provide a plurality of tracks on the scale, side-by-side in parallel. For example, one of the tracks may contain the periodic scale marks themselves, while another contains a reference mark or marks. The reference mark indicates a datum position along the scale, which can be detected by a detector in the readhead. As the readhead passes over the reference mark, it produces a pulse used to reset the counter. This enables the counter to give an indication of absolute position relative to the reference mark. For example, when the encoder is first switched on, it is important that the counter should be reset by the reference mark, otherwise the position indication given will be arbitrary.
When an encoder system is installed, it is necessary to set the readhead up in correct alignment with the scale. In particular, if the readhead is set up with a yaw misalignment, there is an Abbe error between the reference mark and the periodic scale marks.
The present invention, at least in the preferred embodiments, seeks to provide an arrangement for a reference mark and reference mark detector which is less sensitive to yaw misalignment.
The present invention provides an encoder comprising a scale and a readhead which is movable along the scale; the scale comprising a main scale track with a series of periodic scale marks extending along the scale, and two co-operating reference marks associated with a position along the scale and spaced laterally with respect to each other; the readhead including a reference mark detector or detectors for detection of said two reference marks; a combined output being produced from the detection of both said reference marks.
Should the readhead suffer yaw misalignment relative to the scale, the combined reference mark signal can average the Abbe errors from the two reference marks, thereby reducing their effect.
Preferably the two reference marks are located laterally on opposite sides of the main scale track, so that their Abbe errors are equal and opposite and are substantially cancelled when the signals from the reference marks are combined.
A preferred embodiment of the present invention will now be described by way of example, with reference to the accompanying drawings, wherein:
Fig 1 is a schematic isometric view of part of a scale and a readhead of an encoder;
Figs 2,3 and 4 are graphs of detector signals before and after combination; and
Figs 5, 6 and 7 are schematic isometric views of three modifications of the scale and readhead of Fig 1.
Referring to Fig 1, an optical scale 10 has a main scale track 12 with a series of periodically spaced scale marks extending in the longitudinal direction. Laterally on each side of the main scale track 12 are respective reference tracks 14,16. The reference tracks 14,16 each contain one or more reference marks 18,20. The reference marks 18,20, in this embodiment, comprise thin laterally extending substantially specularly reflective lines on a non-reflective background of the reference track.
A readhead is indicated generally at 22 in Fig 1. In practice, the scale 10 and readhead 22 are fixed to respective relatively movable members, such that the readhead moves back and forth along the scale in the longitudinal direction indicated by arrow X. A conventional optical detector 24 in the readhead interacts with the periodic marks of the scale track 12 and produces a pulse train to a counter (not shown) in the conventional manner, indicating the incremental distance travelled along the scale.
To detect the reference marks, the readhead 22 includes two light emitters 26,28, located one on each side of the centre line of the scale. Preferably they are in the form of laterally extending lines, as shown. Each line light emitter 26,28 directs light to a respective reference mark 18,20. The reflective reference marks 18,20 then reflect this light towards a central split
detector 30. The split detector has two photosensitive regions 30A, 30B, each being elongate (similar to the line light emitters 26,28 and the reflective marks 18,20) and extending laterally, spaced in the longitudinal direction from each other.
The two halves 30A, 30B of the split detector 30 produce signals which are taken to a differential amplifier 32. The output of this differential amplifier is indicative of the difference between the two signals, and is taken to a zero crossing detector 34 which provides an output pulse when the difference signal crosses zero. This is taken to the external counter (not shown) in order to reset it. The differential amplifier 32 and zero crossing detector 34 may be provided within the readhead, or in a separate interface circuit between the readhead and the external counter.
In Fig 2, the curve A in the upper graph shows the response of one half 30A of the split detector, as the readhead moves over the reference marks in the longitudinal direction. This is a notional response, indicating the signal resulting from only one of the reference marks and its corresponding line light emitter, isolated from the signal resulting from the other reference mark. The curve B in Fig 2 indicates the corresponding response of the half 30B of the split detector, which of course is spaced from the curve A in the longitudinal direction X. The lower graph in Fig 2 is a curve showing A-B, as output from the differential amplifier 32, and it will be understood that the zero crossing detector 34 produces its output pulse at the point 0 where the curve crosses the X axis.
Fig 3 shows the notional result if the system suffers from yaw misalignment (again considering the signal resulting from just one reference mark) . A yaw error has the effect of shifting the reference signals slightly in the X direction, so that the pulse output by the zero crossing detector 34 no longer occurs at the point 0.
If there is no yaw misalignment, then the combined effect of the responses due to both reference marks
18,20 (and their respective line light emitters 26,28) will be similar to that shown in Fig 2. However, if there is a yaw misalignment, the result is now as shown in Fig 4. The responses A and B due to the two reference marks are shifted in opposite directions by the yaw error, so they combine in the split detector 30 as shown at A1 and B1 respectively. It will be noted that these curves are wider and flatter than in Fig 2, but are centred on the same points as in Fig 2. Importantly, the difference signal A'-B' shown in the lower graph of Fig 4 has a zero crossing point at the point 0 on the X axis; it is not subject to any shift of the type seen in Fig 3.
Thus, the output from the zero crossing detector 34 is insensitive to the yaw misalignment.
Fig 5 shows a modification of the arrangement shown in Fig 1. Here, a single line light emitter 40 is located centrally, in the position of the split detector 30 in Fig 1. This reflects light from the two reference marks 18,20 towards respective split detectors 42,44, located in place of the line light emitters 26,28 shown in Fig 1. The signals from these two split detectors
are combined electronically. Specifically, the signals from each of the A channels are summed in a summing circuit 46 and separately the signals from each of the B channels are summed in a summing circuit 48. The resulting combined signals are then taken to the differential amplifier 32 to provide the A-B signal as previously.
As with the differential amplifier 32 and zero crossing detector 34, the summing circuits 46,48 may be provided within the readhead, or in a separate interface circuit between the readhead and the external counter.
In such an arrangement, it is not essential that there be one single, central light source. Instead, there could be a light source for each of the reference marks 18,20. For each reference mark, the respective light source and split detector may be spaced laterally, as in Fig 6. Alternatively, for each reference mark, the light source and split detector may be spaced longitudinally, as in Fig 7. In the latter case, it will be appreciated that the light source and reference mark are located physically above the reference track 14 or 16 in which the reference mark concerned is provided.
The above description has been of a specularly reflective optical scale. However, the invention can also be applied to a transmissive optical scale, in which the reference tracks 14,16 are generally opaque, with transparent windows forming the reference marks 18,20. The light sources are then located on the other side of the scale relative to the detectors. The
invention can also be applied to diffusely reflective reference marks, e.g. with suitable lens arrangements.
The reference marks do not need to be single lines as seen at 18 and 20. They could instead be, for example, chirped reference marks. Alternatively they could be autocorrelating marks in arrangements such as described in our International Patent Application No. PCT/GB02/00638.
Neither is the invention restricted to optical scales and readheads . For example, a scale could have magnetic reference marks, one on each side of the main scale track. The readhead then has respective magnetic detectors for the reference marks, again on either side of the main scale track, the signals being combined electronically from the two detectors. Yaw insensitivity is again assured by the fact that any yaw misalignment affects the two reference marks in equal and opposite ways. The magnetic detectors could be Hall sensors, reacting to reference marks which are magnetised; or they could be inductive sensors which detect marks made of a ferromagnetic material. Alternatively, the reference marks and their detectors could interact capacitively .
Although the above description has been in respect of a linear scale, it is equally applicable to rotary encoders .
It will be appreciated that there might not be reference tracks 14,16 on the scale itself. Instead, the reference marks 18,20 could be provided on the substrate to which the main scale track 12 is affixed.
Indeed, the main scale track itself could also be provided directly on the substrate.
Particularly in the embodiments of Figs 5, 6 and 7, it will also be appreciated that the reference marks 18,20 need not be spaced at equal distances on opposite sides of the main scale track 12. They could be on opposite sides of the main scale track 12, but spaced by unequal amounts from it. Or they could both be on the same side of the main scale track, again spaced by unequal amounts from it. However, in either of these cases, yawing of the readhead will have unequal effects on the signals from the detectors. For accurate results, therefore, the detected reference mark position will require compensation by electronic or computer processing, depending on the differences in the respective reference mark positions detected by the two detectors, so such arrangements are not preferred.