Encoder Device

Abstract

The position measuring encoder device comprises a scale readable by a readhead and including a series of position features extending along a measurement dimension, the readhead including a polarized light source that emits polarized light towards the scale, a sensor configured to detect optical signals from the position features on the scale that can be used to determine the relative position of the scale and the readhead along the measurement dimension, and a polarization manipulator located in an optical path between the polarized light source and the sensor, the polarization manipulator configured such that the degree of polarization of light output from the polarization manipulator is less than the degree of polarization of polarized light received by the polarization manipulator, regardless of the predominant input polarization direction of the footprint of the polarized light received by the polarization manipulator.

Description

[Technical Field]

[0001] The present invention relates to a position measuring encoder device. [Background technology]

[0002] A position measurement encoder device (hereinafter referred to as "encoder device" or "position encoder") can be used to determine the movement of two relatively movable parts of a device. A position encoder typically comprises a scale and a readhead (one provided on one part of the device and the other provided on another part of the device). The scale may include a series of features that can be read by the readhead to measure its position along the scale (e.g., the scale features may be provided on a substrate that is fixed to one part of the device, or may be integrally formed as part of the device).

[0003] So-called "incremental" position encoders may, for example, function by "counting" positions along the length of the scale, for example from its starting position and / or from a predetermined reference marker on the scale. As will be appreciated, the way in which "counting" the readhead's position is done may vary from encoder device to encoder device. For example, one way is to generate a resultant field, such as an interference fringe field / pattern ("fringe field"), a modulated spot (modulated spot), or an image at the readhead's detector, that changes with relative movement. For example, light from a light source in the readhead may impinge on the scale which diffracts the light into several diffraction orders. The diffraction orders interfere / interact / recombine at the detector to produce the resultant field. Optionally, optical elements (e.g. diffraction gratings and / or lenses) may be provided in the readhead to redirect / deflect the diffraction orders so that they interfere / interact at the detector to produce the resultant field. As the scale and readhead move relative to each other, the resultant field changes. The readhead can record and / or report movement and position by monitoring the resulting field changes (e.g., movement of the interference fringe pattern / fringe field). Such a position encoder is described in U.S. Pat. No. 5,861,953.

[0004] FIG. 1a shows a schematic of the optical operation of such a typical prior art incremental position encoder system. The scale 1000 has a periodic series of features with period “p” that, when illuminated by light, produces a set of diffraction orders 1001, 1002, 1003. It will be understood that references to light herein include visible and invisible light ranging from ultraviolet to infrared. In the particular embodiment described, near-infrared light is used. The diffraction orders 1001, 1002, 1003 are relayed to a sensor 3000 by a relay element 2000 (e.g., a lens, prism, or diffraction grating) (both the sensor and relay element are located within a readhead that is movable relative to the scale 1000). The diffraction orders interfere at the sensor 3000, thereby producing an interference fringe pattern / field (schematically represented by wave 1100) at the sensor 3000 with a period equal to Mp, where M is the magnification of the optical system and p is the scale period. 1a shows the light as being transmitted through the scale, it will be appreciated that the light may be reflected from the scale and therefore may originate from a light source located on the same side of the scale as the sensor. It will be appreciated that typically the light source, relay element 2000 and sensor 3000 will all be provided by a readhead arrangement that is arranged to move relative to the scale 1000.

[0005] Figure 1a is a simplified diagram of the optical situation encountered within an encoder apparatus. In reality, the optical situation shown in Figure 1a is repeated many times along the length of the scale (i.e. across the area illuminated by the light source), producing a longer interference fringe field / pattern at the detector (e.g. as shown schematically in Figure 1b).

[0006] For illustrative purposes, only the 0th and ±1st orders are shown in Figures 1a and 1b. As will be appreciated, higher diffraction orders can be generated and contribute to the formation of the fringe field in sensor 3000, but their intensity, and therefore their contribution to the fringe field, is typically much weaker than the 0th and ±1st diffraction orders (higher orders being progressively weaker in intensity).

[0007] It will be appreciated that if the mark space ratio of the scale (i.e., the ratio of the width of the scale features to the spacing between marks) is exactly 1:1 (as is often the case with amplitude scales in encoder devices), no even diffraction orders (e.g., ±2nd, ±4th diffraction orders) will be produced, and only odd diffraction orders will be produced (e.g., ±3rd, ±5th diffraction orders). In practice, some slight manufacturing errors may mean that the mark space ratio of the scale is not exactly 1:1, and therefore even diffraction orders may be present (although they may be significantly less intense than adjacent odd diffraction orders). It will be appreciated that if the mark space ratio of the scale is intentionally formed so that it is not 1:1, significant even diffraction orders may be present.

[0008] For simplicity of illustration, the ray diagrams in Figures 1a and 1b are shown as transmitted ray diagrams (i.e., light is shown transmitted through each of the scale and optical relay elements), although in practice at least one of these may be reflective.

[0009] As will be appreciated, reference marks may be provided, for example adjacent to and / or embedded within the diffractive features of the scale, to provide predetermined reference positions. Such a position encoder is described in US Patent No. 7,659,992.

[0010] So-called "absolute" position encoders are also known, which can determine the absolute position of a readhead relative to a scale without the need to count from a predetermined position, such as a reference mark or the end position of the scale. Absolute position encoders typically include a scale with unique position data formed thereon along the measuring length of the scale. The data can be in the form of a pseudorandom sequence or discrete code words, for example. By reading this data as the scale reader passes over the scale, the scale reader can determine its absolute position. Examples of absolute position encoders are described in U.S. Pat. No. 7,499,827, U.S. Pat. No. 10,132,657, and U.S. Patent Application Publication No. 2012 / 0072169. Some absolute encoders are known to use incremental scales in conjunction with the absolute scale. Optionally, the absolute scale may retain sufficient periodicity to allow the scale to be used as a periodic incremental scale (e.g., as described in U.S. Pat. No. 7,499,827). In either case, such incremental scales can be used, for example, to fine-tune the determined absolute position. Optionally, after starting up to determine absolute position, such a system can be used to subsequently measure the relative position of the readhead and scale by using the incremental scale to "count" changes in position. Such an incremental scale can be read in the same way as above, for example by analysing the resultant field generated (at a sensor in the readhead) by diffraction orders produced by the scale. Summary of the Invention

[0011] The present invention relates to an improved optical encoder.

[0012] According to a first aspect of the present invention, there is provided a position measuring encoder apparatus comprising: a scale readable by a readhead and including a series of position features extending along a measurement dimension, the readhead including: i) a polarized light source that emits polarized light towards the scale, ii) a sensor configured to detect optical signals from the position features on the scale that can be used to determine the relative position of the scale and the readhead along the measurement dimension, and iii) a polarization manipulator located in an optical path between the polarized light source and the sensor, the polarization manipulator being configured such that the degree of polarization of light output from the polarization manipulator is less than the degree of polarization of light polarized (e.g. emitted from the polarized light source) received by the polarization manipulator, regardless of the dominant input polarization direction of the footprint of light received by the polarization manipulator (or in other words all possible dominant input polarization directions).

[0013] Optical encoders typically employ unpolarized light sources. In fact, at the time of writing, all optical encoders sold by Renishaw plc utilize light-emitting diodes (LEDs) that emit unpolarized light. The present invention provides an alternative optical encoder that utilizes polarized light sources to emit light that illuminates the scale. For purposes of this patent application, polarized light sources refer to light sources that emit light with a degree of polarization of at least 0.2. (In other words, optionally, the degree of polarization of light emitted from a polarized light source is 0.2 or greater.) Similarly, for purposes of this patent application, polarized light refers to light with a degree of polarization of at least 0.2. Light sources can include, for example, laser light sources, such as vertical cavity surface-emitting lasers (VCSELs). As will be appreciated by those skilled in the art, laser light sources, such as VCSELs, emit polarized light that typically has a degree of polarization of at least 0.3 (typically, 0.3 to 0.7).

[0014] Optionally, the light source is a linearly polarized light source, in other words, optionally the light source emits linearly polarized light.

[0015] As will be appreciated, the "degree of polarization" (or, as sometimes referred to herein, "DOP") of light is a simple metric that quantifies how polarized the light is. Perfectly unpolarized light has a DOP of 0 (or 0%), and perfectly polarized light has a DOP of 1 (or 100%). As will be explained herein, the DOP of light can be easily measured experimentally. Light is passed through a linear polarizer, and the exiting light from the linear polarizer strikes a photodetector. The linear polarizer is rotated 180 degrees (as is the rotation of the linear polarizer) and the optical power observed at the photodetector is measured. The optical power at the photodetector will vary in a sinusoidal pattern if the light has polarization. This sinusoidal modulation reveals the DOP. Thus:

[0016]

number

[0017] As is well known, light can be "circularly" polarized or "elliptically" polarized, just as it can be linearly polarized. As will be appreciated, circularly polarized light has a DOP of 0, per the above measurement. Pure linearly polarized light has a DOP of 1, and elliptically polarized light has a DOP between 0 and 1 (inclusive), depending on how elliptical the elliptical polarization is. As will be appreciated, linearly polarized light may not be "pure" or "purely" linearly polarized, and therefore, linearly polarized light may not have a DOP of 1.

[0018] Compared to LEDs, laser light sources can offer the advantage of optical power, which has been found to provide improved measurement performance, particularly reduced jitter due to reduced noise. The inventors have also found that the use of polarization manipulators configured to reduce the degree of polarization of light from polarized light sources, such as lasers, can also be beneficial, as described in more detail below.

[0019] As mentioned above, in the context of the present invention, a polarized light source can be any light source that emits light with a DOP of at least 0.2. However, the DOP of the polarized light source can of course be higher. In fact, the higher the DOP of the polarized light source, the greater the benefits of the polarization manipulator.

[0020] Preferably, the polarization manipulator is configured so that the light emitted therefrom has a DOP of 0.1 or less.

[0021] The polarization manipulator may include a diffuser. Preferably, the polarization manipulator includes a retarder (also known as a "wave plate"), such as a "patterned retarder" (also known as a "non-uniform retarder" or "structured retarder"), as described in more detail below. Thus, the polarization manipulator may be referred to as an optical retarder element.

[0022] It will be understood that the use of the terms "pattern" and "patterned" herein, particularly in relation to polarization manipulators, is not intended to imply the existence of any kind of repeating configuration. Rather, the terms "pattern" and "patterned" are used to refer to a polarization manipulator, e.g., a retarder, having a non-uniform design / shape, and in particular a non-uniform fast axis. As will be explained in more detail below, the (fast axis) design / shape / pattern of a polarization manipulator (e.g., a retarder) can be repeated, but this is not necessarily the case, and in some cases it may be preferable not to repeat.

[0023] The polarization manipulator can include a spatial polarization manipulator or a temporal (or, in other words, "time-based") polarization manipulator. As will be appreciated, in the case of a spatial polarization manipulator, the polarization state of the light is spatially mixed (or, in other words, spatially "scrambled") by the polarization manipulator (i.e., the polarization state is mixed / varied across the footprint of the light exiting the polarization manipulator), whereas in the case of a temporal / time-based polarization manipulator, the polarization state of the light is time / temporally mixed (or, in other words, time / temporally scrambled).

[0024] Optionally, the polarization manipulator is configured so that the light output from the polarization manipulator contains a mixture of polarization directions, regardless of the DOP of the footprint of the light received by the polarization manipulator (e.g., light emitted from a polarized light source) and regardless of the predominant input polarization direction of the footprint. As will be appreciated, in the case of a spatial polarization manipulator, the light output from the polarization manipulator contains a mixture of polarization directions across its footprint / range. In such a case, the polarization manipulator is preferably configured so that the light output from the polarization manipulator has a balanced mixture of polarization directions across its footprint / range (e.g., so that the optical power of the light output therefrom is substantially equal along orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes). Substantially equal may mean that the difference in optical power along any two orthogonal polarization axes is 5% or less, optionally 2% or less, and particularly 1% or less. Preferably, the mixture of polarization directions is substantially evenly distributed across the footprint of the light output from the polarization manipulator. The polarization manipulator may be configured such that, regardless of the DOP of the footprint of the light received by the polarization manipulator and regardless of the predominant input polarization direction of said footprint, the light output from the polarization manipulator has (over its footprint / range) a mixture of at least 4 different polarization directions, more preferably a mixture of at least 8 different polarization directions, optionally a mixture of at least 16 different polarization directions, and particularly preferably a mixture of at least 25 different polarization directions.

[0025] A polarization manipulator (e.g., an optical retarder element) according to the present invention can include a non-uniform fast axis and / or non-uniform retardation characteristic (or waveplate fraction) that varies along at least one axis / dimension. In other words, a polarization manipulator (e.g., an optical retarder element) according to the present invention can include a fast axis and / or retardation characteristic that is non-uniform along at least one axis / dimension. The non-uniform fast axis and / or non-uniform retardation characteristic can vary along a dimension perpendicular to the measurement dimension. The non-uniform fast axis and / or non-uniform retardation characteristic can vary along a dimension parallel to the measurement dimension. The optical retarder element can include a non-uniform fast axis and / or non-uniform retardation characteristic that varies along two orthogonal axes (e.g., can be adjusted to be parallel and perpendicular to the measurement dimension, respectively). However, to avoid adverse diffraction effects, it may be preferable for the non-uniform fast axis and / or non-uniform retardation characteristic to vary only along the dimension perpendicular to the measurement dimension.

[0026] The fast axis and / or retardation characteristics of a polarization manipulator (e.g., optical retarder element) may be varied periodically (i.e., the state of the fast axis may change according to a repeating design / pattern). However, it may be preferable for the fast axis and / or retardation characteristics to be varied non-periodically (i.e., according to a non-repeating design / pattern) (this is particularly true when light passes through the optical retarder element multiple times). Thus, the polarization manipulator (e.g., optical retarder element) may be configured to vary the fast axis and / or retardation characteristics in at least a region of the footprint of the light output from the polarization manipulator / optical retarder element that forms the optical signal at the sensor. , the fast axis and / or retardance (and thus the polarization direction of light emitted therefrom) may be configured to vary non-periodically (in other words, vary according to a non-periodically / non-repeating design / pattern). In particular, the fast axis may vary according to a phase-wrapping linear function of position along at least one dimension of the optical retarder element. Thus, preferred embodiments of non-uniform / patterned retarders include a fast axis that varies according to a phase-wrapping linear function that provides no repetition over at least the length of a region of the footprint of light output from the non-uniform / patterned retarder that forms the optical signal at the sensor.

[0027] In a preferred embodiment of the present invention, the polarization manipulator is configured so that the DOP of the light output from at least a region of the footprint of the light output from the polarization manipulator that forms the optical signal at the sensor (regardless of the predominant input polarization direction of the light received by the polarization manipulator, such light has a DOP of at least 0.2) is smaller than the degree of polarization of the polarized light emitted from the polarized light source, and preferably is less than half of the polarization impinging on the polarization manipulator (e.g., optical retarder element) (or less than half of the polarization emitted from the polarized light source). In particular, the polarization manipulator is preferably configured so that (regardless of the DOP of the light impinging on the polarization manipulator and regardless of the predominant input polarization direction of the light) the degree of polarization of the light emitted from the polarization manipulator on its final pass (e.g., after only one pass or after multiple passes, depending on the embodiment) is less than 0.1, more preferably less than 0.05, e.g., less than 0.02. In particular, preferably, the DOP of the light output from at least a region of the footprint of the light output from the polarization manipulator that forms the optical signal at the sensor is 0.1 or less, for example 0.05 or less, for example 0.02 or less. The polarization manipulator may be configured such that (regardless of the DOP of the light impinging on the polarization manipulator and regardless of the predominant input polarization direction of said light) at least a region of the footprint of the light output from the polarization manipulator that forms the optical signal at the sensor has a substantially balanced mix of different polarization directions (e.g., such that the optical power within said region of the light output from the polarization manipulator is substantially equal along the orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes).

[0028] Optionally, light from the polarized light source is configured to pass through the polarizing manipulator a first time on the way to the scale, and then pass through the polarizing manipulator a second (e.g., last) time after being reflected from the scale. In such a case, preferably the polarizing manipulator is configured such that, after the light has passed through the polarizing manipulator a second (e.g., last) time, the DOP of the light output from at least a region of the footprint of the light output from the polarizing manipulator that forms the optical signal at the sensor is smaller than the DOP of the polarized light emitted from the polarized light source. In particular, preferably the polarizing manipulator is configured such that, after the light has passed through the polarizing manipulator a second time, the DOP of the light output from at least a region of the footprint of the light output from the polarizing manipulator that forms the optical signal at the sensor is 0.1 or less, e.g., 0.05 or less, e.g., 0.02 or less. After the light has passed through the polarization manipulator a second time (e.g., a final time), it may be preferable for at least a region of the footprint of the light output from the polarization manipulator that forms the optical signal at the sensor to have a substantially balanced mix of different polarization directions, such that the optical power within that region of the light output from the polarization manipulator (e.g., an optical retarder element) is substantially equal along the orthogonal polarization axes regardless of the orientation of the orthogonal polarization axes.

[0029] Optionally, in a preferred embodiment of the present invention, the polarization manipulator is configured to conceptually divide the area of ​​the footprint of light output from the polarization manipulator that forms the optical signal at the sensor into a one-dimensional array of n columns of equal width (the width being measured perpendicular to the measuring dimension of the scale), such that the degree of polarization of the light output from each of the n columns is less than the degree of polarization of the polarized light emitted from the polarized light source. Here, n is 2, and the one-dimensional array extends perpendicular to the measuring dimension of the scale. Preferably, the DOP of the light output from each of the n columns is 0.1 or less, e.g., 0.05 or less, e.g., 0.02 or less. Optionally, the optical power of the light output from the polarization manipulator in each of the n columns is substantially equal along the orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes. n can be any integer greater than or equal to 2. Larger values ​​of n result in a greater mixing of polarization states desired from the polarization manipulator, but may be more difficult to implement in practice. Thus, n is at least 2 and is typically an integer less than or equal to 50, e.g., less than or equal to 25. In preferred embodiments, n is 4, more preferably 6, and especially preferably 8. The same statements in this paragraph may also (additionally or alternatively) be true for conceptually dividing the area of ​​the light footprint output from a polarization manipulator that forms the light signal at the sensor into a one-dimensional array of n rows of equal width, where n is 2 and the array extends parallel to the measurement dimension. In embodiments in which light from the polarized light source passes through the polarization manipulator twice, preferably the polarization manipulator is configured such that the requirements of the previous statements in this paragraph are met on both the first and second pass.

[0030] The encoder device may be configured such that in a dimension perpendicular to the measurement dimension; within the area on the sensor illuminated by light from the polarization manipulator; without a polarizer in the optical path between the polarization manipulator and the sensor, the polarization direction of the light from the polarization manipulator varies over a length scale that is at least 5 times, more preferably at least 10 times, particularly preferably at least 15 times, for example at least 20 times, for example at least 25 times smaller than the range of the sensor.

[0031] As mentioned above, the variation can be such that the polarization direction changes periodically (e.g., the polarization state changes according to a design / pattern), but preferably the variation is such that the polarization direction changes according to a non-repeating design / pattern (in other words, preferably the variation is such that the polarization direction changes non-periodically). Excluding any polarizers in the path between the polarization manipulator (e.g., optical retarder element) and the sensor, the actual pitch of the variation in polarization direction as it strikes the sensor will depend on the specific geometry of any given system. Nevertheless, in preferred embodiments, the pitch is 200 μm or less, e.g., 100 μm or less, and preferably 50 μm or less, e.g., 25 μm or less.

[0032] Optionally, in the dimension of the change, the ratio of a) the pitch of the change in the polarization direction of light from the polarization manipulator (e.g., an optical retarder element) to b) the size of the illuminated area is at least 1:25, more preferably at least 1:35.

[0033] In the field of optics, the terms "retarder" and "wave plate" are synonymous. Therefore, another appropriate name for the optical retarder element referred to herein is "optical wave plate element." The polarization manipulator (e.g., optical retarder element) may include a birefringent material. Optionally, the polarization manipulator (e.g., optical retarder element) includes a structure configured to provide birefringence-like properties, such as a metamaterial. For example, it is known that nanocasting lithography can be used to provide wave plates.

[0034] The fast axis and / or retardation characteristics (and thus the polarization of light leaving the optical retarder element) of the polarization manipulator (e.g., optical retarder element) may vary continuously / smoothly. Optionally, the polarization manipulator (e.g., optical retarder element) includes discrete sections or "pixels" of different fast axes and / or different retardation characteristics. In embodiments where the fast axis and / or retardation characteristics (or the polarization of light leaving the polarization manipulator / optical retarder element) vary in only one dimension (e.g., the dimension perpendicular to the measurement dimension), such sections / pixels may be elongated strips or columns. The sections / pixels / strips / columns may be substantially nominally the same size, but this is not necessarily the case. Each section / pixel may, for example, include a discrete waveplate (or "retarder") section. Thus, the polarization manipulator (e.g., optical retarder element) may include a series of discrete waveplate sections. The series can extend in two dimensions, but as described above and in more detail below, the series preferably extends in only one dimension, which may be perpendicular to the measurement dimension. Such waveplate sections can include half-waveplates, quarter-waveplates, or waveplates of different ratios. In accordance with the above description that the retardation characteristics of the polarization manipulator (e.g., optical retarder element) can vary, different sections / pixels can have different waveplate ratios.

[0035] The readhead may further comprise a diffractive lens (e.g. a Fresnel lens) or a refractive lens (e.g. a cylindrical / spherical lens). The lens may preferably be located in the optical path between the light source and the polarization manipulator so that light from the light source only hits / passes the polarization manipulator after passing through the lens. This avoids adverse diffraction effects.

[0036] It may be preferable for the polarization manipulator to be located in the light path before the scale. In other words, it may be preferable for light from the light source to pass through the polarization manipulator before reaching / hitting the scale. However, this does not necessarily have to be the case, and the polarization manipulator can also be located in the light path after the scale. As mentioned above and described in more detail below, the polarization manipulator can be located in the light path before or after the scale (i.e., so that light from the light source passes through the polarization manipulator (e.g., an optical retarder element) twice).

[0037] The readhead may include first and second sensors, each configured to detect an optical signal from a positional feature of the scale. The positional feature of the scale may diffract light into a plurality of diffraction orders. At least one diffraction order may have a polarization state that is different from the polarization state of at least one other diffraction order. The apparatus may be configured such that the signal sensed by the first sensor is formed from a diffraction order composition that is different from the diffraction order composition of the signal sensed by the second sensor, depending on the polarization states of the diffraction orders.

[0038] The series of positional features of the scale can diffract light into multiple diffraction orders. The encoder device can be configured so that at least one diffraction order has a polarization state that is different from the polarization state of at least one other diffraction order. The polarization states of the diffraction orders can be imparted / encoded by the scale. Thus, the scale can include / be a diffraction order encoder. For example, the scale can include a birefringent scale or a holographic scale, such as those described in U.S. Patent Application Publication No. 2003 / 0141441. The birefringent scale or holographic scale includes a polarization hologram recorded on the scale such that the polarization directions of the ±1 diffraction orders are rotated 90° relative to the 0 diffraction order.

[0039] In a preferred embodiment, the readhead includes a diffraction order encoder that encodes at least one diffraction order produced by the series of positional features on the scale in a polarization state that differs from the polarization state of at least one other diffraction order produced by the series of positional features on the scale. Providing a diffraction order encoder in the readhead can be simpler and cheaper than providing a holographic scale. Further details of such a diffraction order encoder are provided below.

[0040] The readhead may be configured to at least partially filter at least one diffraction order (e.g., the zeroth diffraction order) based on its polarization state so as to at least partially attenuate its effect on the generation of a signal sensed by at least one sensor of the readhead.

[0041] The readhead may include first and second sensors, each configured to detect an optical signal from a positional feature of the scale. The readhead may be configured such that the signal sensed by the first sensor is formed from a different diffraction order composition than the signal sensed by the second sensor, depending on the polarization state of the diffraction orders. Such an encoder device may provide improved performance over existing encoder devices. In particular, the signals sensed by the different sensors may be tailored to provide improved / optimized signals for each of the different sensors. For example, it may be beneficial to suppress the zeroth diffraction order from contributing to the signal sensed by the first sensor (which may be, for example, an incremental sensor), but not to do so for the second sensor (which may be, for example, a reference mark sensor or absolute sensor). Therefore, since it is not desirable to completely block such diffraction orders from propagating toward the sensors of the readhead, the present invention instead facilitates selective / tailored suppression of diffraction orders for different sensors of the readhead based on their polarization states. For example, the zeroth diffraction order may have a polarization state different from the polarization states of the ±1st diffraction orders.

[0042] The diffraction order composition may vary from sensor to sensor, in the sense that the degree to which different diffraction orders contribute / influence the signal sensed by the sensor varies for different sensors. For example, both the signal sensed by a first sensor and the signal sensed by a second sensor may be formed from the zeroth diffraction order and ±1st diffraction orders, but due to their polarization states, the relative influence of the zeroth diffraction order and ±1st diffraction orders on the formation of the signal sensed by the first sensor is different from the relative influence of the zeroth diffraction order and ±1st diffraction orders on the formation of the signal sensed by the second sensor. For example, the device may be configured so that the influence of the zeroth diffraction order on the signal sensed by the first sensor is 50% less than the influence of the zeroth diffraction order on the signal sensed by the second sensor.

[0043] While some benefit can be obtained from only partially attenuating the influence of a particular diffraction order (e.g., the zeroth or ±1st diffraction orders) on generating the signal formed at the sensor, for example by attenuating the influence of the particular diffraction order on signal generation by at least 50%, as described above, it is often preferable for the influence of a particular diffraction order on generating the signal sensed by the sensor to be substantially completely attenuated, for example by at least 90%, for example by at least 95%, for example by at least 98%. In particularly preferred embodiments, the readhead is configured so that one or more particular diffraction orders (e.g., the zeroth or ±1st diffraction orders) have substantially no effect on generating the signal sensed by one of the first and second sensors, while their effect on generating the signal sensed by the other sensor remains unaffected. In fact, it has been found that the present invention provides greatest benefit when selected diffraction orders are substantially completely suppressed for at least one of the sensors while remaining substantially unsuppressed for the other sensor.

[0044] Thus, the apparatus may be configured such that the signals sensed by the first and second sensors may be formed / composed of different subsets of diffraction orders based on their polarization states. For example, in one embodiment, the signal sensed by the first sensor may be formed of one or more diffraction orders other than the zeroth diffraction order (i.e., the zeroth diffraction order does not contribute to the signal sensed by the first sensor), while the zeroth diffraction order contributes (optionally together with other diffraction orders) to the signal sensed by the second sensor.

[0045] In the field of encoder devices, the zeroth and ±1st diffraction orders often have the strongest influence on the signal formed by / sensed by a sensor. Therefore, in preferred embodiments of the present invention, the zeroth diffraction order has a polarization state that is different from the polarization states of the ±1st diffraction orders. This allows selective control of the extent to which the zeroth and / or ±1st diffraction orders influence or contribute to the signal sensed by the sensor. While the +1st diffraction order can be encoded to have a polarization state that is different from the polarization state of the −1st diffraction order, it is typically beneficial to configure the readhead so that they have the same polarization state.

[0046] The device may be configured (e.g., by a polarizer integrated into the sensor) such that the sensor (e.g., the first sensor and / or the second sensor) is essentially at least partially blind to certain polarization states (and thus essentially at least partially blind to one or more diffraction orders). Thus, the composition of the signal as sensed by the sensor is determined by the extent to which the sensor is at least partially blind to polarization states. In preferred embodiments, the encoder device may include a sensor filter configured to at least partially (e.g., substantially) filter light based on the polarization state before the light hits the sensor. For example, in embodiments in which the readhead includes a first sensor and a second sensor, the readhead may include a first sensor filter configured to at least partially (e.g., substantially) filter light based on the polarization state before the light hits the first sensor and / or a second sensor filter configured to at least partially (e.g., substantially) filter light based on the polarization state before the light hits the second sensor. Thus, diffraction orders may be selectively at least partially (e.g., substantially) filtered by the first sensor filter and / or the second sensor filter based on their polarization states. Thus, light impinging on the first sensor may be filtered differently than light impinging on the second sensor. If both a first sensor filter and a second sensor filter are provided, they may be configured to filter at least in part (e.g., substantially) different diffraction orders based on their polarization states. Thus, the composition of the signal as sensed by the first and / or second sensors may be determined / controlled by the filters.

[0047] In particularly preferred embodiments, the sensor filter (e.g., the first sensor filter) substantially prevents light of all zeroth diffraction orders from reaching the (e.g., first) sensor (in other words, substantially filters the zeroth diffraction order). In other particularly preferred embodiments, the sensor filter (e.g., the second sensor filter) substantially prevents light of all ±1st diffraction orders from reaching the (e.g., second) sensor (in other words, substantially filters the ±1st diffraction orders). As explained in more detail below, in preferred embodiments, diffraction orders greater than the ±1st diffraction orders are substantially suppressed / filtered / prevented from reaching the sensor. Thus, in preferred embodiments, the signal formed at the (e.g., first) sensor is formed primarily (or exclusively) from the ±1st diffraction orders. In other preferred embodiments, the signal formed at the (e.g., second) sensor is formed primarily (or exclusively) from the zeroth diffraction order.

[0048] The above-described diffraction order encoder can include at least one polarizer element configured to encode at least one diffraction order in a polarization state different from the polarization state of at least one other diffraction order. In particularly preferred embodiments, the diffraction order encoder comprises a first polarizer element configured to encode the zeroth diffraction order in a first polarization state and one or more additional (e.g., second and third) polarizer elements configured to encode the ±1st diffraction orders in a second polarization state different from the first polarization state. Optionally, the polarizer element includes a polarizer configured to polarize (e.g., increase the degree of polarization) at least one diffraction order. Optionally, the polarizer element includes at least one polarization manipulator, such as a waveplate or retarder, configured to change the polarization state of at least one diffraction order, e.g., to rotate the polarization direction of the at least one diffraction order. In such cases, the diffraction orders may need to be polarized before the diffraction order encoder. For example, a polarizer may be positioned in the optical path before the waveplate and after the polarization manipulator. Using a polarization manipulator to reduce the DOP of light emitted from a polarized light source, but then re-polarizing the light again before it reaches the diffraction order encoder, can be advantageous over simply allowing the raw polarized light from the polarized light source to hit the waveplate of the diffraction order encoder. This is because the polarization direction of the polarized light source may be unpredictable / unstable. To provide a known, constant polarization of the light, a polarizer can be used after the polarization manipulator but before the diffraction order encoder.

[0049] In embodiments in which the readhead includes a diffraction order encoder, the light source and the diffraction order encoder may be provided on a single mounting member, in other words on a single substrate, for example a glass substrate.

[0050] The readhead may include at least one optical relay element for relaying light, e.g., the diffraction orders, from the scale to the sensors (e.g., the first and second sensors). For example, the readhead may include at least one refractive and / or diffractive optical relay element, e.g., the diffraction orders, from the scale to the sensors (e.g., the first and second sensors). Suitable optical relay elements include lenses and / or diffraction gratings. A diffraction order encoder may be positioned to interact with one or more diffraction orders before or after the optical relay element (e.g., the diffraction orders). Optionally, the diffraction order encoder and the optical relay element (e.g., the diffraction orders) may be one and the same (e.g., at least one optical relay element may include a diffraction order encoder integrated therein).

[0051] Optionally, the device is configured so that the diffraction orders converge to corresponding / respective spots (or "different convergence points") in the optical path before the sensors (e.g., the first and second sensors). For example, there may be a spot / convergence point for each diffraction order, e.g., a spot / convergence point for the 0th diffraction order, a spot / convergence point for the +1st diffraction order, a spot / convergence point for the -1st diffraction order, etc. Such spots may be located in the focal plane (focal plane of the lens) of an optical relay element for relaying the diffraction orders. A diffraction order encoder may be located substantially at these convergence points. For example, the diffraction order encoder may be located in the focal plane of the optical relay element, e.g., the focal plane of the optical relay element. The spots may be formed in a conjugate plane of the light source. Thus, the spots may be images of the light source.

[0052] The polarization states of diffraction orders greater than ±1 diffraction orders can be selectively controlled to selectively control how much they affect or contribute to the signal sensed by the first and / or second sensors. Such diffraction orders can include, for example, ±3 diffraction orders and / or ±5 diffraction orders. Alternatively, in some embodiments, many of which are preferred embodiments, diffraction orders greater than ±1 diffraction orders are simply prevented / stopped / blocked from reaching the sensors (e.g., the first and second sensors) so as not to affect the signal generated thereat. Such diffraction orders can be stopped by absorption, deflection, scattering, and / or reflection. For example, an opaque material can be placed at the spot / convergence point for selecting all diffraction orders greater than ±1 diffraction orders at the conjugate plane of the light source to absorb and block such diffraction orders. Optionally, the diffraction order encoder is configured to encode diffraction orders greater than the ±1 diffraction orders in polarization states such that at the sensor (e.g., at the first and second sensors) they do not interact / interfere with the 0th and / or ±1st diffraction orders, or such that they are at least partially (e.g., substantially) filtered by an appropriate filter (e.g., filtered by the first sensor filter and / or the second sensor filter) before reaching the sensor.

[0053] The scale may be illuminated with collimated light (and optionally the diffraction orders from the scale may themselves be collimated). Optionally, the readhead includes an optical collimator element for collimating the light from the light source. Optionally, the same optical collimator element (for collimating the light from the light source) and the optical relay element for relaying the light (e.g. the diffraction orders) towards the sensors (e.g. the first and second sensors) are one and the same / optical element. Thus, optionally the same optical element is used for collimating the light from the light source and focusing the diffraction orders onto corresponding / respective spots.

[0054] The apparatus may be configured such that the sensors (e.g., the first and second sensors) lie substantially in conjugate planes of the scale. The position measurement encoder apparatus may therefore be described as an imaging encoder apparatus, and an image (or pseudo-image) of the scale is formed by the sensors (e.g., the first and second sensors).

[0055] Preferably, the scale comprises what is commonly referred to as an amplitude or "Ronchi" scale (see phase scale). As will be appreciated, in an amplitude or "Ronchi" scale, features are configured to control the amplitude of light reflected (or transmitted in transmissive scale embodiments) towards the readhead (particularly towards its first and second sensors), for example by selectively absorbing, scattering, and / or reflecting light. In contrast, a phase scale is configured to control the phase of light reflected (or transmitted in transmissive scale embodiments) towards the readhead (particularly towards its first and second sensors), for example by controlling scale features at different depths to less than a fraction of the wavelength of light. Typically, an amplitude scale produces a significant zeroth diffraction order, along with significant ±1st diffraction orders (the higher ±odd diffraction orders which decrease in intensity). This is in contrast to a phase scale, which does not produce a zeroth diffraction order.

[0056] Optionally, the period of the scale is 40 μm or less, preferably 20 μm or less, such as 10 μm or less, for example 8 μm or less.

[0057] Preferably, the scale has a feature space (or "mark" space) ratio of 1:1, so in other words the ratio of the width of the scale features to the spacing of the scale is 1:1.

[0058] The scale may be a transmissive scale. Preferably, the scale is a reflective scale. Therefore, the light source and the first and second sensors of the readhead are preferably located on the same side of the scale.

[0059] Optionally, the position measurement encoder device is a one-grating encoder system, and the scale includes only a diffraction grating in the optical path between the light source and the first and second sensors.

[0060] The series of position features of the scale may be provided in at least one track (a "scale track"). The scale may include one or more scale tracks.

[0061] The signal generated by the sensor may include an incremental position signal, e.g., an interference fringe pattern or a modulated spot. Thus, the sensor may include an incremental position sensor. The period of the interference fringes may be Mp / 2, where M is the magnification of the encoder's optics and p is the period of the scale (which may be achieved when the zeroth diffraction order is substantially eliminated). The scale may include an incremental scale track including a series of periodic features that define the incremental scale track. One or more reference marks may be provided within, embedded in, and / or located adjacent to the incremental scale track. Such reference marks may include optical reference marks. Thus, optionally, the signal generated by the sensor may include a reference mark signal. In embodiments in which the readhead includes first and second sensors, the second sensor may be configured to detect the reference mark signal generated by the reference mark.

[0062] Optionally, the signal generated by the sensor includes an absolute position signal. Thus, the sensor may include an absolute position sensor. Thus, the scale may include an absolute scale track including a series of features defining an absolute scale track. As will be appreciated, an absolute scale track differs from an incremental scale track (with or without reference marks) in that its features define a series of unique positions along the length of the scale. It may be read by the readhead such that the relative positions of the readhead and scale can be determined (e.g., upon start-up) at any position along the scale without requiring movement to a reference position (e.g., reference mark). Examples of absolute scales include those described in U.S. Patent Nos. 7,499,827 and 5,279,044. The scale may include a separate incremental track and an absolute scale track. Optionally, the incremental and absolute scale features are combined into one track. For example, the absolute scale features may be superimposed on periodic incremental scale features. In embodiments in which the readhead includes first and second sensors, the first sensor may be configured to detect an incremental scale / signal (and therefore may also be referred to as an incremental sensor) and the second sensor may be configured to detect an absolute scale / signal (and therefore may also be referred to as an absolute sensor), or vice versa. As is known, and as described in the prior art referenced above in this paragraph, absolute position information can be encoded in the scale track by omitting selected position features from a periodic series of other position features.

[0063] Optionally, the scale includes first and second series of position features readable by the readhead. The first series of position features can generate a first set of diffraction orders, and the second series of position features can generate a second set of diffraction orders (which generate signals detected by the first and second sensors). These first and second sets of diffraction orders can be superimposed / spatially overlapping with each other. The position measurement encoder device can be configured such that only one of the first and second sets of diffraction orders has a diffraction order with a different optical state. For example, the position encoder device can be configured such that at least one diffraction order (e.g., the zeroth diffraction order) of the first set of diffraction orders has an optical state that is different from the optical state of at least one other diffraction order (e.g., the ±1st diffraction orders) of the first set of diffraction orders, while some diffraction orders (e.g., at least the zeroth diffraction order and the ±1st diffraction orders) of the second set of diffraction orders can have the same optical state. Alternatively, the position measurement encoder apparatus may be configured such that at least one diffraction order of the first set has an optical state that is different from the optical state of at least one other diffraction order, and such that at least one diffraction order of the second set has an optical state that is different from the optical state of at least one other diffraction order. In any event, it will be appreciated that the first and second sets of diffraction orders may be filtered differently based on their optical states.

[0064] The first and / or second series of position features may be periodic (in other words, the scale may include a first periodic series of position features and a second periodic series of position features). The period of the first series of position features may be different from the period of the second series of position features. For example, the scale may include a scale track including a first series of position features having a first period (e.g., a relatively fine period) and a second series of position features having a second period (e.g., a relatively coarse period, i.e., coarser than the period of the first period). The first and / or second series of position features may be provided on the same track, i.e., embedded / superimposed on each other.

[0065] As will be appreciated, the output of the first sensor (and the output of the second sensor, if present) can provide one or more signals that can be used to indicate relative position information (with respect to the readhead and scale). As will be explained in more detail later in this specification, such position information can be incremental or absolute position information. Such position information can be index (also known as "reference" or "datum") position information. The position information can be linear or angular position information (e.g., in the case of a rotary encoder device). The readhead can output raw, unprocessed signals from the sensors. Optionally, the readhead can process the signals from the sensors and output one or more signals derived from the signals from the sensors. As with a position encoder, the readhead can output one or more signals, for example, quadrature signals (e.g., sine and cosine signals) that vary with changes in the relative position of the scale and readhead (i.e., vary with relative motion). The readhead can output analog or digital signals that represent the relative position of the scale and readhead. For example, the quadrature signals can be analog quadrature signals or digital quadrature signals. Optionally, the readhead may maintain and output an incremental "count" of the position of the readhead and the scale (e.g., it may count from an index position). Optionally, the readhead may output a digital codeword representing the relative position of the scale and readhead. For example, the readhead may output an absolute digital codeword representing the absolute relative position of the scale and readhead. It will also be appreciated that a controller may use the output of the readhead to determine how to control an apparatus / device to which the encoder apparatus is attached.

[0066] The position measuring encoder device may be a linear encoder device or a rotary encoder device, and therefore the scale may be a linear scale or a rotary scale (in which case the scale may be a ring scale or a disk scale).

[0067] The position measuring encoder device may be a linear encoder device or a rotary encoder device, and therefore the scale may be a linear scale or a rotary scale (in which case the scale may be a ring scale or a disk scale).

[0068] As will be appreciated, the output of the first sensor (and the output of the second sensor, if present) can provide one or more signals that can be used to indicate relative position information (with respect to the readhead and scale). As will be explained in more detail later in this specification, such position information can be incremental or absolute position information. Such position information can be index (also known as "reference" or "datum") position information. The position information can be linear or angular position information (e.g., in the case of a rotary encoder device). The readhead can output raw, unprocessed signals from the sensor. Optionally, the readhead can process the signals from the sensor and output one or more signals derived from the signals from the sensor. As with a position encoder, the readhead can output sine and cosine signals (commonly referred to as "quadrature" signals) that vary with changes in the relative position of the scale and readhead (i.e., vary with relative motion). The readhead can output analog or digital signals that represent the relative position of the scale and readhead. For example, the quadrature signals can be analog quadrature signals or digital quadrature signals. Optionally, the readhead may maintain and output an incremental "count" of the position of the readhead and the scale (e.g., it may count from an index position). Optionally, the readhead may output a digital codeword representing the relative position of the scale and readhead. For example, the readhead may output an absolute digital codeword representing the absolute relative position of the scale and readhead. It will also be appreciated that a controller may use the output of the readhead to determine how to control an apparatus / device to which the encoder apparatus is attached.

[0069] According to a second aspect of the present invention, there is provided a position measuring encoder apparatus comprising: a scale including a series of location features extending across the measurement dimension; The read head and Equipped with The read head is i. a polarized light source that emits polarized light toward the scale; ii. a sensor configured to detect an optical signal from a position feature of the scale; and iii. an optical retarder element located in an optical path between the polarized light source and the sensor, the optical retarder element having non-uniform fast axis and / or retardation characteristics along at least one dimension, configured such that light output from the optical retarder element (over its footprint / extent) contains a mix of polarization directions regardless of the predominant input polarization direction of the footprint of polarized light impinging on the optical retarder element; Thus, different sections / portions of light output from the optical retarder element will have different polarization directions. The features described above in relation to the first aspect of the invention are also applicable to this second aspect of the invention. [Brief explanation of the drawings]

[0070] Hereinafter, an embodiment of the present invention will be described by way of example with reference to the drawings.

[0071] [Figure 1a] FIG. 1 is a schematic diagram illustrating the optical operation of a typical prior art position encoder system. [Figure 1b] FIG. 1 is a schematic diagram illustrating the optical operation of a typical prior art position encoder system. [Figure 2] 1 shows an incremental scale and readhead of an encoder device according to the present invention. [Figure 3] 3 shows schematically various optical components located within the readhead of FIG. 2 according to a first embodiment of the invention; [Figure 4(a)] 6 illustrates an exemplary electrical grid sensor for use in the readhead of FIGS. [Figure 4(b)] 4(a) shows a schematic diagram of bands of light with different polarizations impinging on the incremental sensor of FIG. [Figure 5(a)]4 shows an exemplary reference mark signal from a reference mark photodiode of the readhead of FIGS. 2 and 3; [Figure 5(b)] The resulting differential signal is shown. [Figure 6(a)] 1 shows an exemplary polarization manipulator for use in a readhead according to the present invention; [Figure 6(b)] 6(a) shows four different (i-iv) exemplary orientations of the orthogonal polarization axes (v, h) and a graph of the optical power along the orthogonal polarization axes of the light output from the polarization manipulator of FIG. 6(a) in the light. [Figure 7] 10 illustrates schematically an encoder device according to a second embodiment of the present invention. [Figure 8] FIG. 8 is a schematic optical diagram of the encoder device of FIG. 7. [Figure 9] FIG. 8 is a schematic optical diagram of the encoder device of FIG. 7. [Figure 10] 10 is a graph showing the effect of the zeroth order term on the visibility of interference fringes impinging on the sensor as the distance between the scale and readhead is changed. [Figure 11] 10 is a schematic diagram illustrating an encoder device according to a third embodiment of the present invention. [Figure 12] FIG. 12 is a schematic optical diagram of the encoder device of FIG. [Figure 13] FIG. 12 is a schematic optical diagram of the encoder device of FIG. [Figure 14] 1 shows an exemplary polarization manipulator for use in a readhead according to the present invention; [Figure 15a] 10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 15b] 10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 15c] 10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 16a]10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 16b] 10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 16c] 10 is a graph illustrating the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detector plane of the readhead. [Figure 17] 10A and 10B show schematic diagrams of enlarged views of small sections of incremental scale features; [Figure 18] 10 shows a schematic optical diagram of an encoder arrangement according to another embodiment of the present invention; [Figure 19] 10 shows a schematic diagram of an absolute scale and a readhead of an encoder device according to another embodiment of the present invention. [Figure 20] 10 shows a schematic diagram of an absolute scale and a readhead of an encoder device according to another embodiment of the present invention. [Figure 21] 10 shows a scale and a readhead of an encoder device according to another embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0072] Referring to Figure 2, an encoder device 2 according to the present invention comprises a scale 4 and a readhead 6. The readhead 6 is movable in the X dimension relative to the scale 4 (although this may be in the opposite direction, or indeed both may be movable). For example, the scale 4 may be attached to a stationary part of a machine (not shown) and the readhead 6 may be attached to a movable part of the machine (not shown). The scale 4 may take many different forms including, for example, a linear scale as shown, or a rotary scale (e.g. provided on a ring with scale features provided on its periphery, or provided on a disk with scale features provided on its flat surface).

[0073] In the described embodiment, the scale 4 is an amplitude scale (in that it controls the amplitude of light that passes towards and away from the scale) and includes a substrate having incremental features 10 in the form of periodic dark / relatively non-reflective lines made on an otherwise relatively reflective substrate, so as to provide relative reflection between the incremental features 10 and the scale. Of course, the scale incremental features 10 could be made in other ways, for example by forming relatively reflective lines on an otherwise relatively non-reflective substrate, or even by forming both relatively reflective and non-reflective lines on a substrate. It will also be appreciated that the incremental features 10 could be provided in other ways, for example they could be provided in the form of reflective surfaces or lines that reflect light towards and away from the readhead. In the described embodiment, the scale substrate is metal, but it will be appreciated that other materials could be used, such as glass, for example.

[0074] In an alternative embodiment, the scale 4 may be a phase scale in which peaks and pits in the scale modulate the phase of light leaving the scale towards the readhead. However, as will become apparent throughout the remainder of this specification, the present invention has been found to be particularly useful when used in conjunction with an amplitude scale rather than a phase scale due to the presence of the zeroth diffraction order produced by an amplitude scale (whereas the zeroth diffraction order does not tend to be produced by a phase scale).

[0075] The incremental features 10 form an incremental scale that facilitates measurement along the X axis (the "measurement dimension"). In the illustrated embodiment, a reference mark 11 is provided, and in particular is embedded within the incremental features 10. In the described embodiment, the reference mark comprises a single continuous block formed by the omission of multiple continuous reflective lines. As will be appreciated, the reference mark does not necessarily have to be embedded within the incremental scale track; instead, it may be provided in a separate track. In other embodiments, the reference mark is non-optical, or, for example, not provided at all.

[0076] Figure 3 shows various optical components located within the readhead 6 (for clarity, the body has been omitted and the relative sizes and positions of components are not to scale). In this embodiment the readhead includes a polarised light source 12, a polarisation manipulator 14 (which in this embodiment is an optical retarder element, also referred to herein as a "non-uniform retarder" or "patterned retarder", as will be explained in more detail below), a diffraction grating 16, an optical imaging element 18, a primary position information sensor 20 and a secondary position information sensor 22.

[0077] As previously mentioned herein, the use of the terms "pattern" and "patterned" herein, particularly in relation to polarization manipulators, is not intended to imply that there is any repeating configuration. Rather, the terms "pattern" and "patterned" are used to refer to a polarization manipulator, e.g., a retarder, having a non-uniform design / formation.

[0078] Except for the use of polarized light source 12 and optical retarder element 14, the configuration and operation of this embodiment of the encoder is substantially identical to that described in WO2005124282. Thus, with optical retarder element 14 set aside, light from light source 12 illuminates area 24 of scale 4. Due to the periodic arrangement of incremental features 10, light reflected by incremental features 10 is diffracted into diffraction orders (i.e., in the same manner as described above in connection with FIGS. 1a and 1b). The diffracted light strikes diffraction grating 16, where the light is diffracted into further diffraction orders and recombines at primary position information sensor 20 (hereinafter referred to as "incremental sensor" 20) to form an interference fringe pattern (or "fringe field") thereon. Movement of readhead 6 relative to scale 4 causes movement of the interference fringes relative to incremental sensor 20, thus generating an up / down count that allows for measurement of the variation.

[0079] In the described embodiment, incremental sensor 20 is in the form of an electrical grid, or in other words, an optical sensor array including two or more sets of interdigitated / interlaced / interleaved light-sensitive sensor elements (also referred to herein as "photodetectors" or "fingers"). Each set can, for example, detect a different phase of the interference fringe pattern / field (schematically represented by wave 1100) in incremental sensor 20. FIG. 4(a) shows a portion of an electrical grid, where four sets of photodiode fingers / photodiodes (A, B, C, and D) are interdigitated / interleaved to form an array of sensor elements extending along the length "L" of the sensor.

[0080] The outputs from each finger / photodiode in a set are combined to provide a single output, resulting in four channel outputs: A', B', C', and D'. These outputs are then used to obtain quadrature SIN and COS signals. In particular, A'-C' are used to provide a first signal (SIN), and B'-D' are used to provide a second signal (COS) that is 90 degrees out of phase with the first signal. In certain embodiments, the electrical grid includes four sets of photodiodes providing four channels A', B', C', and D', although this need not be the case. For example, the electrical grid may include two sets of photodiodes providing only two channels, A' and B'.

[0081] It will be appreciated that other types of sensors can be used in place of the electrical gratings described above. For example, in embodiments where a modulated spot is created by the readhead optics instead of an interference fringe pattern, a bulk sensor photodiode can be used to detect the intensity of the modulated spot (e.g., as described in U.S. Pat. No. 4,776,701). In systems that generate modulated spots, similar to systems that generate interference fringe patterns, encoding the 0th and ±1st diffraction orders so that they do not interfere with each other at the sensor effectively doubles the resolution of the encoder system (in that the intensity of the spot modulates at a frequency of p / 2). Similarly, encoding diffraction orders higher than the ±1st diffraction orders in an optical state different from that of the ±1st diffraction orders reduces / eliminates harmonics in the modulated spot intensity.

[0082] The polarized light source 12 emits polarized light, i.e., light having at least a dominant direction of polarization, if not only one. Polarized light sources can include linearly polarized light sources, circularly polarized light sources, or elliptically polarized light sources. In this embodiment, the polarized light source 12 comprises a laser light source, in particular a VCSEL, which emits linearly polarized light with a dominant direction of polarization. It will be appreciated that the polarization direction and / or degree of polarization (DOP) of light from laser light sources such as VCSELs can vary unpredictably from light source to light source. Furthermore, the polarization direction and / or DOP of light from laser light sources such as VCSELs can change during operation of the readhead 6 due to changes in the operating environment, for example, changes in temperature or applied current (which can be changed by servo control). Typically, the DOP of the type of VCSEL used in the described embodiment is in the range of 0.3 to 0.7.

[0083] As the readhead 6 passes the reference mark 11, changes in the intensity of light reflected by the scale 4 are imaged by the optical imager lens 18 onto a secondary position information sensor 22 (hereinafter referred to as the "reference mark sensor"). The secondary position information sensor 22 includes a split detector including first and second photodiodes 22a, 22b onto which light from the scale 4 is imaged. FIG. 5(a) schematically shows the outputs (S1, S2) from the first and second photodiodes 22a, 22b as the readhead 6 moves along the scale 4 and passes the reference mark 11. The signals S1 and S2 from the first and second photodiodes 22a, 22b are subtracted to form a difference signal S3, shown in FIG. 5(b). By identifying where the falling edge of signal S3 crosses a known threshold in the illustrated reading direction, the position of the reference mark can be determined. It will be appreciated that although the present embodiment uses a split detector to detect the reference mark, other types of detectors and / or other techniques for identifying the presence of a reference mark may be used. The optical imaging element 18 shown in Figure 3 is a Fresnel zone plate. However, other types of optical imaging elements having the same optical function may be used, such as a refractive lens. As described in WO2005124282, the reference mark sensor 22 may be partially or completely embedded within the incremental sensor 20 (similarly, the optical imaging element 18 may be partially or completely embedded within the diffraction grating 16).

[0084] As mentioned above, the VCSEL 12 emits polarized light. The inventors have identified that the polarization of the light used to illuminate the scale can, in some circumstances, adversely affect the optical signal formed by the sensors 20, 22. For example, high-pitch scales (e.g., feature pitches less than 1 μm) can exhibit birefringence and therefore be sensitive to the polarization of the light impinging on the scale. Furthermore, holographic scales can also be polarization-sensitive. Therefore, reducing the DOP of the light from the VCSEL can be beneficial. This can make calibration / mapping-out errors caused by polarization direction difficult, if not impossible, especially when the polarization direction of the light source is not stable (e.g., as is the case with VCSELs) or can change over time. While laser light sources (e.g., edge-emitting laser light sources) with a fixed polarization direction are known, the present invention enables the use of light sources with unstable polarization directions, such as VCSELs, which can be cheaper, more reliable, easier to package, consume less power, and / or provide a more symmetrical spatial profile than edge-emitting laser light sources.

[0085] 3, in accordance with the present invention, an optical retarder element 14 that reduces the DOP of the light from the VCSEL 12 is positioned in the optical path between the VCSEL 12 and the sensors 20, 22. In the particular embodiment described, the optical retarder element 14 is located in the optical path before the light interacts with a diffractive optical element in the optical system, such as the scale 4.

[0086] In accordance with the present invention, the optical retarder element 14 comprises a half-wave plate having a non-uniform fast axis along at least one dimension. An exemplary optical retarder element 14 is shown in more detail in FIG. 6( a). In this embodiment, the optical retarder element 14 comprises a non-uniform fast axis along the Y dimension, which is orthogonal to the measuring dimension (X) of the encoder (and thus may be described as a “non-uniform retarder” or “patterned retarder”). In particular, the optical retarder element 14 comprises an array of discrete elongated half-wave plate pixels / columns 19, the array extending in the Y dimension and the elongated length of the pixels / columns extending in the X dimension. Thus, the array / series of pixels extends perpendicular to the measuring dimension (X) of the scale 4. Thus, in this embodiment, the optical retarder element (also referred to as a non-uniform / patterned retarder) 14 does not include structure in the measuring dimension (X). This may be advantageous to prevent the non-uniform / patterned retarder 14 from being a source of diffraction of light in the X dimension, which may otherwise adversely affect the generation of the optical signal at the sensors 20, 22. However, as will be appreciated, in other embodiments / applications, such diffraction may be acceptable, and thus the non-uniform / patterned retarder may include a structure in the X dimension, for example, it may be arranged such that an array / series of retarder pixels extend along the Y dimension, or the non-uniform / patterned retarder may include a two-dimensional array of retarder pixels.

[0087] As shown in FIG. 6( a), the half-wave plate pixels 19 are arranged with alternating fast axes at 0° and 45°. Such a configuration ensures that light output from the optical retarder element 14 has a mix of polarization directions across its extent / footprint (in the Y dimension) such that the DOP of the light leaving the optical retarder element 14 is less than the DOP of the light impinging on the optical retarder element 14, regardless of the predominant input polarization direction of the footprint 21 of light impinging on the optical retarder element. In particular, in this embodiment, the light output from the optical retarder element 14 always includes alternating sections / portions / columns (in the Y dimension) of light having two different polarization directions that are orthogonal to each other, regardless of the predominant input polarization direction. Furthermore, in the described embodiment, the light output from the optical retarder element 14 includes a balanced mix of polarization directions (e.g., in this embodiment, there are substantially equal amounts of the two different orthogonal polarization directions).

[0088] Thus, the optical retarder element 14 is configured so that the optical power (e.g., milliwatts, “mW”) of the light within the optical footprint 21 output from the optical retarder element is substantially equal along the orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes (thus, the DOP of the optical footprint 21 output from the optical retarder element is close to 0, e.g., 0.1 or less, and preferably 0.02 or less). This can be beneficial for reasons of balance and consistency in encoder performance. For example, if the optical power along one polarization axis is substantially greater than another orthogonal polarization axis, especially if the respective powers change with changes in input polarization direction, the effects of polarization on encoder performance may not be as suppressed as in a configuration in which the optical power of the light along the orthogonal polarization axes is substantially the same.

[0089] Referring to FIG. 6(b), four different (i-iv) exemplary orientations of orthogonal polarization axes (v, h) are shown, along with a graph of the optical power along the orthogonal polarization axes of light output from the optical retarder element within the optical footprint 21. As shown, due to the substantially balanced mix of polarization directions output by the optical retarder element (also referred to as a non-uniform / patterned retarder) 14 within the optical footprint 21, for any pair of orthogonal polarization axes, the optical power along each polarization axis is substantially the same, regardless of the orientation of the orthogonal polarization axes. In other words, within the optical retarder element output footprint 21, if the light output from the optical retarder element is polarized along a first polarization axis, the optical power of that polarization will be substantially the same as the optical power of the light if it were instead polarized along a second polarization axis that is orthogonal to the first polarization axis.

[0090] The optical retarder element is configured such that, within at least a region of a footprint of light output from the optical retarder element that forms the optical signal at the sensor, regardless of the predominant input polarization direction of the light impinging on the optical retarder element, the difference in optical power along the orthogonal polarization axis of the output beam is at least half that of the input beam, regardless of the orientation of the orthogonal polarization axis.

[0091] In the embodiments described herein, substantially equal (or "substantially the same") optical power along orthogonal polarization axes means that the difference in optical power along the orthogonal polarization axes varies by no more than 2% of the total power. Thus, in such cases, the DOP of the light output from the optical retarder element within the optical footprint 21 is no more than 0.02 or 2%. However, it will be appreciated that in other embodiments, such tight tolerances may not be required, and thus substantially equal (or "substantially the same") optical power of the light along orthogonal polarization axes may mean that the difference in optical power along the orthogonal polarization axes varies by no more than 5% of the total power, e.g., no more than 10% of the total power, e.g., no more than 20% of the total power. In other words, the optical retarder element may be configured such that the DOP of the light output from the optical retarder element within the optical footprint 21 is no more than 0.05 (or 5%), e.g., no more than 0.1 (or 10%), e.g., no more than 0.2 (or 20%).

[0092] As shown in FIG. 6( a), the pitch p of the half-wave plate pixels 19 is substantially smaller than the width of the optical retarder element 14 in the Y dimension. In particular, although the optical retarder element 14 may contain only two pixels (so that the first half of the light output therefrom has one polarization direction and the second half of the light output therefrom has a different, orthogonal polarization direction), it has been found to be beneficial to provide the optical retarder element 14 with substantially more half-wave plate pixels 19 to spread different polarization directions (rather than the same or similar polarization directions all grouped to one side) across the footprint of the light output from the optical retarder element 14, and therefore across the scale features on the scale 4. This can be beneficial for a number of reasons, including providing greater freedom for where a sensor can be placed along the Y dimension and still receive the benefits of the present invention. In the present embodiment, this is achieved by providing the optical retarder element (also referred to as a non-uniform / patterned retarder) 14 with multiple half-wave plate pixels 19 of alternating fast axes. In particular, the pitch p1 of the half-wave plate pixels 19 (in the dimension along which the series of pixels extends, in this embodiment along the Y dimension) and the width W of the optical retarder element 14 r The ratio of is at least 1:4, more preferably at least 1:10, particularly preferably at least 1:20, for example at least 1:30.

[0093] It should be noted that in this embodiment, the arrangement of half-wave plate pixels 19 applies equally to region 21a of the footprint of the light output from the optical retarder element forming the optical signal at incremental sensor 20 and region 21b of the footprint of the light output from the optical retarder element forming the optical signal at reference mark sensor 22. Thus, it is ensured that the light impinging on incremental sensor 20 and reference mark sensor 22 contains a balanced mix of polarization directions. Therefore, in the described embodiment, graphs (i-iv) of FIG. 6(b) apply equally to region 21a of the footprint of the light output from the optical retarder element forming the optical signal at incremental sensor 20 and region 21b of the footprint of the light output from the optical retarder element forming the optical signal at reference mark sensor 22.

[0094] Furthermore, in the described embodiment, the optical retarder element 14 is configured such that the different polarization directions output within region 21a are substantially evenly distributed in the Y dimension (i.e., perpendicular to the measuring dimension X), such that, to conceptually divide the region into a one-dimensional array of two equal-width columns (21aa and 21ab) extending parallel to the Y dimension, the optical power along the orthogonal polarization axes of the light output from the optical retarder element within each of the two columns is substantially equal, regardless of the orientation of the orthogonal polarization axes. Note that this is also true for region 21b of the footprint of the light output from the optical retarder element forming the optical signal at reference mark sensor 22 (e.g., to conceptually divide the region into a one-dimensional array of two equal-width columns 21ba and 21bb extending parallel to the Y dimension, the optical power along the orthogonal polarization axes of the light output from the optical retarder element within each of the n columns is substantially equal, regardless of the orientation of the orthogonal polarization axes). Thus, graphs (i-iv) of FIG. 6(b) (and to some extent the description of the DOP of the light output from the optical retarder elements within the optical footprint 21) apply equally to each row 21aa and 21ab of the region 21a of the optical footprint output from the optical retarder elements forming the optical signal at the incremental sensor 20, and equally to each row 21ba and 21bb of the region 21b of the optical footprint output from the optical retarder elements forming the optical signal at the reference mark sensor 22. Such a configuration can help suppress errors that might otherwise be caused by imperfections in the scale features along their length. Additionally, uniform illumination across the sensor can be beneficial. For example, the sensor may be shaped / windowed / weighted to improve the signal output by the sensor, as described, for example, in U.S. Pat. No. 1,067,0431, and uniform illumination may be required to maintain the benefits / effects of the shaping / windowing / weighting.

[0095] FIG. 4(b) (which is similar to FIG. 4(a) in that it shows a portion of incremental sensor 20, in this case in the form of an electrical grid) schematically illustrates bands 23 of light having different polarizations impinging on incremental sensor 20 (the polarization of each band of light impinging on the sensor is indicated by an arrow). As conceptually illustrated, in the dimension perpendicular to the measurement dimension (i.e., in the dimension Y perpendicular to measurement dimension X), the pitch p2 of the variations in polarization direction of light impinging on incremental sensor 20 is s 4(b) is substantially smaller than . This can be beneficial because such an arrangement can help suppress errors that may be caused by imperfections in the scale features along their length by spreading out different polarization directions over the length of the features. In the particular embodiment shown, the ratio of a) the pitch of the change in polarization direction of the light from the optical retarder element to b) the size of the illuminated area is 1:7 to 1:8. In some preferred embodiments, the ratio can be much larger, for example, at least 1:25, but is shown at a much lower ratio in FIG. 4(b) for ease and clarity of illustration.

[0096] An alternative embodiment of an encoder device 102 according to the present invention will now be described with reference to Figures 7 to 9. This embodiment shares some parts that are the same as the embodiment of Figure 3, and similar parts share the same reference numerals.

[0097] Differences from the embodiment of FIG. 3 include omitting the diffraction grating and Fresnel zone plate of the embodiment of FIG. 3 and providing imaging optics in the form of a refractive lens 109. Additionally, a diffraction order encoder 116 is provided. The light source 12 and the diffraction order encoder 116 are provided on an opaque substrate 115 in substantially the same plane. As will be appreciated, the substrate 115 need not actually be opaque, but being opaque means that other diffraction orders that may be useful are blocked. Further differences include that, unlike the continuous reference mark 11 of FIG. 3, the reference mark 111 of this embodiment comprises a patterned reference mark, which in this embodiment comprises one thick dark band and two thin dark bands, each separated by at least one incremental feature between them along the measurement direction. The term “patterned” in the expression “patterned reference mark” is not used to mean that the reference mark has a repeating design / form, but rather is used to refer to the reference mark having a non-uniform design / form. Of course, a "patterned reference mark" may or may not have a repeating shape / design. Reference mark sensor 122 includes two sets of photodiodes 122a, 122b, 122c of corresponding placement and shape / size located on opposite sides of incremental sensor 20. Each set of photodiodes 122a, 122b, 122c of reference mark sensor 122 is configured such that only when readhead 106 and reference mark 111 are aligned will an image (or "pseudo image," described in more detail below) of the pattern of reference mark 111 strike and align / correlate with the pattern of the three photodiodes 122a, 122b, 122c of reference mark sensor 122, thereby causing a significant and sudden change in the intensity of light impinging thereon. In this embodiment, where reference mark 111 includes dark bands, the intensity of light received by reference mark sensor 122 drops significantly when readhead 106 and reference mark 111 are aligned.Signal processing electronics and / or software downstream of the reference mark sensor 122 may be configured to identify such changes in the intensity of light received at the reference mark sensor 122 and output a signal indicative of the presence of a reference mark. As far as the downstream signal processing electronics and / or software are concerned, the two sets of correspondingly positioned photodiodes 122a, 122b, 122c function as one photodiode.

[0098] The optical configuration of the encoder apparatus 102 of FIG. 7 will now be described in more detail with reference to FIGS. 8 and 9. Light from the VCSEL 12 is emitted towards the scale 4 through an optical retarder element (also referred to as a non-uniform / patterned retarder) 14 (configured in the same way as described above in connection with FIGS. 2-6) so that the light output therefrom contains two orthogonal polarization directions. The light output from the non-uniform / patterned retarder 14 is relayed to the scale 4 via a lens 109 (through a clear / transparent window 117 in the readhead 106). The VCSEL 12 is positioned at a focal length f of the lens 109 so that the light from the VCSEL 12 is collimated by the lens 109 when it strikes the scale 4. The light from the VCSEL 12 illuminates the footprint 24 on the scale 104 (see FIG. 7). In FIGS. 8 and 9, thick dashed lines schematically indicate light traveling towards the scale, and thin solid lines indicate light reflected from / by the scale.

[0099] The lens 109 relays light reflected by the scale 104 to the incremental sensor 20 and the reference mark sensor 122 .

[0100] As will be appreciated, due to the well-known natural phenomenon of diffraction, light reflected by the scale 104 is diffracted due to the presence of scale features thereon. In the case of the periodic incremental features 10, the light reflected thereby is diffracted into distinguishable diffraction orders. As discussed above in connection with FIGS. 1a and 1b, light is diffracted in the X dimension into the 0th and ±1st diffraction orders. Higher diffraction orders (e.g., ±3rd and ±5th orders) also exist but are not shown because they have substantially lower intensities than the 0th and ±1st diffraction orders. In either case, they are blocked by the opaque substrate 115 from proceeding to either the incremental sensor 20 or the reference mark sensor 122.

[0101] The diffracted orders are incident on lens 109, which focuses the corresponding diffracted orders to form spots 150 at the back focal plane fp of lens 109, each spot formed from a respective diffracted order. Spots 150 are images of the light source at the point where the back focal plane fp is a conjugate plane to the plane in which the light source is located.

[0102] In this embodiment, the diffraction order encoder 116 is coexistent with the spot 150. In particular, in this embodiment, the diffraction order encoder 116 includes a polarizing filter 116a (e.g., with a "vertical" polarization axis) that is coincident with the zeroth diffraction order spot 150a, and first and second non-polarizing transparent regions 116b and 116c that are coincident with the ±1st diffraction order spots 150b, 150c. Thus, the diffraction order encoder 116 encodes the zeroth diffraction order with only vertical polarization, while the polarizations of the ±1st diffraction orders remain unaffected (and thus include orthogonal polarizations, such as the output from the non-uniform / patterned retarder 14).

[0103] The zeroth and ±1st diffraction orders propagate toward the incremental sensor 20 and the reference mark sensor 122. The primary sensor filter 120 is located in the optical path of the diffraction orders traveling toward the incremental sensor 20. The primary sensor filter 120 includes a polarizing filter with a "horizontal" polarization axis, i.e., an axis orthogonal to the polarization axis of the first polarizing filter 116a. Thus, the zeroth diffraction order light is blocked from reaching the primary position information sensor. In contrast, the ±1st diffraction orders (despite being "horizontally" polarized during processing) are able to pass through the primary sensor filter 120 and thereby impinge on the incremental sensor 20. In particular, the ±1st diffraction orders propagate toward the incremental sensor 20, interact (constructively and destructively interfere), and form an interference fringe pattern (or "fringe field") that impinges on the incremental sensor 20.

[0104] Blocking the zeroth diffraction order results in many improvements to the interference fringe pattern impinging on incremental sensor 20. In particular, fringe pattern visibility is improved. Additionally, removal of the zeroth diffraction order means that the interference fringe pattern produced by incremental sensor 20 has a period equal to Mp / 2 (half that of a system that does not block the zeroth diffraction order). Thus, blocking the zeroth diffraction order effectively doubles the system resolution.

[0105] Figure 10 illustrates another benefit of removing the zeroth diffraction order. In particular, Figure 10 shows how blocking the zeroth diffraction order affects the visibility of fringe patterns impinging on the incremental sensor 20 as the distance between the scale 4 and the readhead 106 (commonly known as the "rideheight") varies. As shown, if the zeroth diffraction order contributes fully to the fringe field, the visibility of the fringe field will be reduced by p 2 The modulation is done with a period of p / λ, where p is the scale period and λ is the wavelength of light. If the zeroth order is eliminated, there is no longer any modulation. This is beneficial because the encoder device that produces the zeroth diffraction order can have a rideheight tolerance that is independent of the scale period.

[0106] So, in summary, blocking the zeroth diffraction order produced by the scale improves the absolute fringe visibility of the fringe field impinging on the incremental sensor 20, effectively doubling the system resolution and significantly improving the rideheight tolerance of the readhead.

[0107] Furthermore, if diffraction orders higher than the first are also blocked (as is the case in this embodiment with opaque substrate 115), higher harmonics are removed from the fringe field, providing a purer interference fringe pattern. It will be appreciated that in alternative embodiments, diffraction orders higher than the first need not be blocked; for example, substrate 115 may be transparent (in which case first and second non-polarizing transparent regions 116b and 116c may not be clearly distinguishable).

[0108] In an alternative embodiment shown in Figures 7-9, the primary sensor filter 120 can be omitted. In this case, both the 0th and ±1st diffraction orders will impinge on the incremental sensor 20. However, because they have different polarization states, they will not interfere with each other. The effect this has on fringe visibility is shown in Figure 10. As shown, similar to the embodiment that prevents the 0th diffraction order from reaching the incremental sensor, the p of the fringe visibility due to changes in ride height is 2 λ / λ modulation is eliminated. However, the absolute visibility of the fringes is less than in the embodiment that prevents the 0th diffraction order from reaching the incremental sensor.

[0109] In the described embodiment, there is no corresponding filter in front of the reference mark sensor 122. Thus, the zeroth and ±1st diffraction orders all impinge on the reference mark sensor 122. Depending on various factors (discussed in more detail below), it may be preferable in some circumstances to not place a corresponding filter in front of the reference mark sensor. Indeed, if the presence of both the zeroth and ±1st diffraction orders impinging on the reference mark sensor does not have a deleterious effect on the optical signal, e.g., an “image” or “pseudo image” (discussed below), formed at the reference mark sensor 122, then there is no advantage to having a corresponding filter in front of the reference mark sensor 122 (indeed, doing so may be detrimental due to reduced photometric performance).

[0110] Therefore, as described above, the signal sensed by incremental sensor 20 is formed from only the ±1st diffraction orders, while the signal sensed by reference mark sensor 122 is formed from the 0th and ±1st diffraction orders. Therefore, the signal sensed by incremental sensor 20 is formed from a diffraction order composition that is different from the diffraction order composition of the signal sensed by reference mark sensor 122.

[0111] One thing to note is that in this embodiment, the zeroth and ±1st diffraction orders are unequally attenuated because the zeroth order term is polarized by the first polarizer 116a. If necessary, this unequally attenuated state can be avoided by replacing the first and second non-polarizing transparent regions 116b, 116c with second and third polarizing filters 116b, 116c having "horizontal" polarization axes (i.e., polarization axes orthogonal to the polarization axis of the first polarizing filter 116a). In such an embodiment, the zeroth and ±1st diffraction orders impinging on the reference mark sensor 122 are equally attenuated, which may result in a better quality image impinging on the reference mark sensor. Furthermore, the ±1st diffraction orders still pass through the primary sensor filter 120 and interfere to form an interference fringe pattern on the incremental sensor 20. In any event, it will be understood that the light impinging on the incremental sensor 20 is filtered differently than the light impinging on the reference mark sensor 122.

[0112] Light impinging on the incremental sensor 20 and reference mark sensor 122 can be described as an "image" of the scale 4, in that the incremental sensor 20 and reference mark sensor 122 lie at a detection plane dp, which in the embodiment shown and described is coplanar with the conjugate plane cp of the scale surface sp (e.g. light from point A on the scale is imaged to point A' on the sensor, light from point B on the scale is imaged to point B' on the sensor and light from point C on the scale is imaged to point C' on the sensor). However, it will be appreciated that suppressing / blocking selective diffraction orders such that they do not reach the sensor will affect the "image" as seen at the conjugate plane cp. For example, the removal of the zeroth diffraction order (and diffraction orders greater than the first diffraction order because the opaque substrate 115 blocks them) means that the "image" at the conjugate plane c p when it strikes the incremental sensor 20 has a purer sinusoidal waveform rather than a square waveform, and so the image on the sensor is not in fact a "true" or "perfect" image of the scale 4 (which might rather be referred to as a pseudo-image of the scale). As will be appreciated, due to the presence of the zeroth order term in the reference mark sensor 122, the "image" of the scale 4 on the reference mark sensor 122 will be affected to a lesser extent, particularly when the zeroth and ±1st diffraction orders are polarized by the diffraction order encoder 116 so that they both strike the reference mark sensor 122 to the same extent.

[0113] As can be seen from the above discussion, without the optical retarder element (also referred to as a non-uniform / patterned retarder) 14, the light from the VCSEL light source 12 reaching the diffraction order encoder 116 would be polarized. In particular, without the optical retarder element (non-uniform / patterned retarder) 14, what strikes the sensor 20, 122 would vary depending on the predominant polarization direction emitted by the VCSEL 12. In the worst case, this could mean that the incremental sensor 20 sees no light at all (e.g., if the VCSEL light source 12 emits only vertically polarized light, as is the case in this embodiment, in which case the ±1st order terms blocked by the primary sensor filter 120 would be vertically polarized). Thus, the presence of the non-uniform / patterned retarder 14 ensures that the incremental sensor 20 receives a proper signal regardless of the polarization direction of the VCSEL light source 12. Indeed, the particular design of the non-uniform / patterned retarder 14 in the described embodiments will ensure that the optical power of the signal impinging on the incremental sensor 20 is the same regardless of the input polarization direction commanded / output by the VCSEL 12. In particular, in the described embodiments, the non-uniform / patterned retarder 14 will ensure that the DOP of the light leaving it is substantially 0 (0%) (e.g., less than 1%, in particular less than 0.05%), regardless of polarization direction, even though the DOP of the light from the VCSEL 12 is 1 (100%).

[0114] 2-6, it may be beneficial to design the optical retarder element (also referred to as the non-uniform / patterned retarder) 14 so that, within at least the region of the footprint of the light output from the non-uniform / patterned retarder 14 that will form the optical signal at the incremental sensor 20, the different polarization directions are substantially evenly distributed along the Y dimension, thereby conceptually dividing the region into a one-dimensional array of two equal-width columns (e.g., 21aa and 21ab) extending parallel to the Y dimension, such that the optical power of the orthogonal polarization axes of the light output from the optical retarder element within each of the two columns is substantially equal, regardless of the orientation of the orthogonal polarization axes. As previously mentioned, this is advantageous because it helps suppress errors that might otherwise be caused by imperfections in the scale features along their length. This is because such a configuration ensures that the signal at the incremental sensor 20 is formed from light coming from different points along the length of the scale features (i.e., in the Y dimension orthogonal to the measurement dimension X).

[0115] As will be appreciated, due to the presence of the primary sensor filter 120, only horizontally polarized light will strike the incremental sensor 20. Thus, unlike the embodiment of Figure 4(b), the light striking the sensor 20 will not have bands 23 of alternating polarization states. Nevertheless, to facilitate a uniform distribution of polarization directions and help suppress errors that might otherwise be caused by imperfections in the scale features along their length, an optical retarder element (also referred to as a non-uniform / patterned retarder) 14 may be used. In the Y dimension, In the area on the sensor 20 illuminated by light from the non-uniform / patterned retarder 14, In the absence of a polarizer in the optical path between the optical retarder element and the sensor, it may be preferred that the polarization direction of the light from the optical retarder element varies over a length scale that is at least 5 times smaller than the range of the sensor, more preferably at least 10 times, particularly preferably at least 15 times, such as at least 20 times, for example at least 25 times smaller.

[0116] An alternative embodiment of an encoder apparatus 202 according to the present invention is shown in Figures 11, 12 and 13. This embodiment shares some parts that are the same as the embodiment of Figures 7 to 9, and similar parts share the same reference numerals.

[0117] In this embodiment, the optical retarder element (also referred to as a non-uniform / patterned retarder) 214 is different from the previous embodiment. First, in this embodiment, the non-uniform / patterned retarder 214 is located in the optical path between the lens 109 and the scale 104 (i.e., between the lens 109 and the window 117 of the readhead 206). The inventors have discovered that if the non-uniform / patterned retarder 214 is located in front of the lens 109, before the light hits the scale 104, it may be beneficial to locate the non-uniform / patterned retarder 214 after the lens 109 due to the detrimental diffractive optical effects that exist. Such detrimental diffractive optical effects may adversely affect the formation of the optical signal formed in the sensor 20, 122.

[0118] The diffraction order encoder 216 of this embodiment differs from the embodiment of Figures 7-9. In this embodiment, the diffraction order encoder 216 includes first, second, and third polarizing filters 216a, 216b, 216c that are aligned with the 0th, +1st, and -1st diffraction order spots 150a, 150b, 150c, respectively, to polarize the 0th, +1st, and -1st diffraction orders. The polarization axes of the second and third polarizing filters 216b, 216c are configured in the same direction as each other so that the ±1st diffraction orders are polarized by the second and third polarizing filters 216b, 216c and have the same (e.g., "horizontal") polarization direction as each other. The polarization axis of the first polarizing filter 216a is configured in a different, preferably orthogonal, direction from the polarization axes of the second and third polarizing filters 216b, 216c, such that the zeroth diffraction order is polarized in a polarization direction different (e.g., "perpendicular") from the polarization direction of the ±1st diffraction orders. Thus, the diffraction order encoder 216 encodes the +1st and -1st diffraction orders in a first polarization direction and encodes the zeroth diffraction order in a second polarization direction different from the first polarization direction. As shown, the first, second, and third polarizing filters 216a, 216b, 216c are spaced apart so that they do not directly contact each other, although this need not be the case.

[0119] The polarized diffraction orders propagate toward the incremental sensor 20 and the reference mark sensor 122. In the same manner as in the embodiment described above in connection with the embodiment of FIG. 7, the primary sensor filter 120 is positioned in the optical path of the diffraction orders traveling toward the incremental sensor 20. The primary sensor filter 120 includes a polarizing filter having a polarization axis parallel to the polarization axes of the second and third polarizing filters 216b, 216c (and thus orthogonal to the polarization axis of the first polarizing filter 216a). Thus, the zeroth diffraction order light is blocked from reaching the incremental sensor 20. In contrast, the ±1st diffraction orders pass through the primary sensor filter 120, thereby allowing them to impinge on the incremental sensor 20. In particular, the ±1st diffraction orders propagate toward the incremental sensor 20 and interact (constructively and destructively interfere) to form fringe fields that impinge on the incremental sensor 20. The benefits of preventing the zeroth diffraction order from contributing to the optical signal impinging on the incremental sensor 20 were described above in connection with the embodiment of FIG. 7.

[0120] 7 embodiment, the secondary sensor filter 223 is located in the optical path of the diffraction orders traveling toward the reference mark sensor 122. In particular, the secondary sensor filter 223 comprises a polarizing filter having a polarization axis parallel to the polarization axis of the first polarizing filter 216a (and therefore orthogonal to the polarization axes of the second and third polarizing filters 216b, 216c). The secondary sensor filter 223 will therefore block the first diffraction orders and prevent them from contributing to the optical signal impinging on the reference mark sensor 122. Therefore, only the zeroth order diffracted light will reach the reference mark sensor 122 and contribute to the optical signal impinging on the reference mark sensor 122. Such an arrangement has been found to be advantageous, particularly in embodiments with incremental features located within the reference mark.

[0121] For example, with reference to Figures 15 and 16, the effect of filtering the ±1st diffraction orders on an image (or "pseudo image") at the detection plane of the readhead is shown. In particular, Figures 15a, 15b, and 15c respectively show images of the scale reconstructed at the detection plane dp / conjugate plane cp of the readhead from both the 0th and ±1st diffraction orders when the readhead is positioned a) at the readhead's nominal ride height, b) +75 μm from the nominal ride height, and c) +150 μm from the nominal ride height. As shown, at the nominal ride height, the image of the reference mark (highlighted by the circle in Figures 15a-15c) is good, but as the readhead moves away from the nominal ride height, the image of the reference mark becomes substantially corrupted. This has been found to be a particular problem when the reference mark contains periodic incremental features, but can also be an issue (albeit to a lesser extent) for reference marks that do not contain incremental features. Such image corruption causes the signal output by the reference mark sensor 122 to become wider and less distinct as the readhead passes over the reference mark, which can cause problems with reference mark reliability (e.g. the signal is too weak to detect the reference mark) and / or reference mark repeatability (e.g. the reference mark signal is too broad and not repeatable within one incremental period).

[0122] Figures 16a, 16b, and 16c show images of the scale reconstructed from only the zeroth diffraction order when the readhead is positioned a) at the readhead's nominal ride height, b) +75 μm from the nominal ride height, and c) +150 μm from the nominal ride height, respectively. As shown in Figure 16, the image of the reference mark reconstructed at the readhead's detection plane dp / conjugate plane cp (highlighted by a circle in Figures 16a-16c) maintains good structure and similarity to the reference mark, providing a clear and strong reference mark signal from the reference mark sensor 122, even at ride heights away from the nominal ride height. We have therefore found it beneficial to provide a secondary sensor filter 223 in front of the reference mark sensor 122 that filters out the ±1st diffraction orders so that they do not contribute to the image of the scale reconstructed at the readhead's detection plane dp / conjugate plane cp. In particular, such a configuration provides the readhead with better ride height tolerance.

[0123] It will be appreciated that, if desired, the reference mark sensor 122 can include a split detector similar to that of Figure 3, where the photodiodes 122a, 122b, 122c are duplicated and laterally offset in the X dimension, and the outputs from the duplicate photodiodes are connected to form a second reference mark output which can be used to provide a difference signal as in Figure 5(b) from which the presence of a reference mark can be detected.

[0124] Therefore, as described above, the signal sensed by incremental sensor 20 is formed from only the ±1st diffraction orders, while the signal sensed by reference mark sensor 122 is formed from only the 0th diffraction order. Thus, the signal sensed by incremental sensor 20 is formed from a diffraction order composition that is different from the diffraction order composition of the signal sensed by reference mark sensor 122.

[0125] As shown in FIGS. 12 and 13 , in contrast to the embodiments of FIGS. 7–9 , light from the VCSEL 12 passes through the optical retarder element (also referred to as a non-uniform / patterned retarder) 214 twice: once toward the scale 104 and again from the scale back toward the sensor (20, 122). In practice, the degree of overlap at the non-uniform / patterned retarder 214 of the light beams to and from the scale is likely to be greater than that shown schematically in FIG. 12 . It has been found that moving the non-uniform / patterned retarder 214 so that it is positioned after the lens 109 (with respect to the light emitted from the VCSEL 12 traveling toward the scale 104) avoids the adverse diffractive optical effects described above, but that such a double pass of light through the non-uniform / patterned retarder 214 presents new challenges for the designer of the encoder device. In fact, the polarization direction of the light returning through the non-uniform / patterned retarder 214 the second time will be changed. The inventors have found that this can cause many sections / portions of the light to be "de-polarized", resulting in the light output from the non-uniform / patterned retarder 214 after passing through the non-uniform / patterned retarder 214 twice having an undesirable non-uniform mix of polarization directions, which in turn can lead to the light after two passes still having its dominant polarization direction (or in other words still having a DOP that does not result in a substantial / meaningful difference in the DOP of the beam emitted from the light source compared to the DOP of the beam after passing through the non-uniform / patterned retarder 214 twice, for example, greater than 0.2 or 20%).

[0126] The inventors have determined that in a double-pass system, an optical retarder element (also referred to as a non-uniform / patterned retarder) 114 having half-wave plate pixels 109 arranged with their different fast axes in a periodic arrangement (i.e., like the non-uniform / patterned retarder 14 of the embodiment of Figures 2-10) results in a significantly unbalanced distribution of polarization directions propagating toward the diffraction order encoder 116, leading to adverse results. The inventors have identified that for a double-pass system, the mix of polarization directions leaving the non-uniform / patterned retarder 14 twice, and therefore propagating toward the diffraction order encoder 116, can be more evenly balanced by providing a non-uniform / patterned retarder 114 having half-wave plate pixels 109 arranged with their different fast axes in a non-periodic arrangement. While the inventors have found that random non-periodic arrangements typically provide an improvement over periodic arrangements, they have found that some non-periodic arrangements of fast axes are superior to others in providing a balanced mix of polarization directions propagating towards the diffraction order encoder 116, regardless of the dominant input polarization direction of the footprint of light initially striking the non-uniform / patterned retarder 114 from the VCSEL 12. As discussed above, a substantially balanced mix of polarization directions within any given region of light output from the non-uniform / patterned retarder means that the optical power of light output from the non-uniform / patterned retarder in that region is substantially equal along the orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes.

[0127] 14 shows an example optical retarder element (also referred to as a non-uniform / patterned retarder) suitable for use in a double-pass configuration and capable of providing a well-balanced mix of polarization directions propagating towards the sensor 20, 122. The following table lists the fast axis directions of each of the 60 pixels that make up the non-uniform / patterned retarder 114:

[0128] [Table 1]

[0129] Thus, as shown, in this embodiment, the non-uniform / patterned retarder 114 includes a plurality (in this embodiment, 60) of half-wave plate pixels 119 having three or more different fast axis directions, which are not arranged periodically.

[0130] Of course, the above arrangement of the fast axes is not the only arrangement that can provide a substantially balanced mix of polarization directions output from the non-uniform / patterned retarder 114 after the light has passed twice. Other arrangements are possible and can be selected by the designer of the optical encoder system. When designing such a non-uniform / patterned retarder, the designer can consider if the initial polarization state of the light is given by:

[0131]

number

[0132] where θ is the angle of linear polarization, and the polarization state after one pass through the non-uniform / patterned retarder (assuming a half-wave plate) is given by:

[0133]

number

[0134] where α(x,y) is the fast axis direction of the non-uniform / patterned retarder as a function of x and y. In the particular embodiments described, the fast axis of the non-uniform / patterned retarder varies in only one dimension, i.e., the Y dimension, so in these cases α(x,y) can be written simply as α(y). After reflection from the scale and passing through another non-uniform / patterned retarder, the polarization state becomes:

[0135]

number

[0136] Here, β(x',y') (which, as mentioned above, can be written as β(y') for the specific embodiment described) is the fast axis direction of the second non-uniform / patterned retarder at point (x',y'). In this embodiment, the two non-uniform / patterned retarders, designated α and β, are concatenated and are the same device. When α = β, the second pass undoes the first pass, giving the original polarization state. This means that for each photon to undo its polarization change after two passes, there must be a position offset between the first and second pass. Also, if α and therefore β vary linearly, then β - α is the same everywhere, and we see that there is no polarization scrambling after the second pass, only pure rotation; this ignores diffraction effects. If the pixel width and linear function are such that diffraction is favored, the light will spread out enough so that the light exiting α(x,y) (or α(y)) does not all enter β(x',y') (or β(y')). This results in scrambling and not just pure rotation.

[0137] There are several desirable aspects of a preferred system that can help guide the selection of the fast axis of the half-wave plate pixels 119, including i) minimal reduction in the visibility of the interference fringe pattern formed on the incremental sensor 20, ii) consistent optical power on the sensor 20, 122, and iii) uniform sampling of the scale 104.

[0138] i) Regarding the visibility of the interference fringe pattern, slight diffraction along the measurement direction of the scale (X) can significantly reduce the visibility of the interference fringes. A simple way to mitigate this problem is to only change the fast axis direction in the dimension perpendicular to the measurement direction of the scale (Y), thereby reducing / eliminating the possibility of diffraction in the measurement dimension (X). This also has the benefit of simplifying the design and reducing the parameter space for design optimization.

[0139] ii) For consistent optical power on the sensor 20, 122, as the dominant input polarization direction changes, the mix of polarization states of the light output from the non-uniform / patterned retarder 114 also changes. This has the ramifications of changing the ratio and distribution of optical power in the orthogonal polarization states after the second pass through the non-uniform / patterned retarder 114. By evenly mixing the polarization states on a length scale much smaller than the footprint of the light as it strikes the sensor, consistency of optical power can be achieved. Thus, as discussed above in connection with the embodiment of Figures 7-10, the non-uniform / patterned retarder 114: - in the Y dimension, in the area on the sensor (e.g., sensor 20) illuminated by light from the non-uniform / patterned retarder 114; - In the absence of a polarizer in the optical path between the non-uniform / patterned retarder 114 and the sensor, it may be preferable to configure the polarization direction of the light from the optical retarder element to vary over a length scale that is at least 5 times smaller than the range of the sensor, more preferably at least 10 times smaller, particularly preferably at least 15 times smaller, such as at least 20 times smaller, for example at least 25 times smaller.

[0140] iii) With regard to uniform sampling of the scale, the scale is unlikely to be a perfect diffraction grating and will therefore introduce errors in line placement and shape. For example, FIG. 17 shows that, at a microscopic level, incremental scale features 10 are not perfectly straight and uniform along their length. Therefore, as discussed above in connection with other embodiments, it may be beneficial to ensure that the light forming the signal at the sensor comes from different points on the scale along the Y dimension, which in this embodiment means that due to the polarization diffraction order encoder 216 and polarizers 120, 223 in front of the sensors 20, 122, it may be beneficial for the light leaving the non-uniform / patterned retarder 114 the second time to have an appropriate spread / distribution of different polarization directions in the Y dimension. Otherwise, the signal formed at each sensor 20, 122 may be formed from light coming substantially from one region / strip of the scale (in the Y dimension). As discussed above in connection with other embodiments, a suitable (e.g., uniform) distribution of different polarization directions in the Y dimension can be a distribution that ensures that the optical power of the light output from the optical retarder elements within each of the two rows is substantially equal along the orthogonal polarization axes, regardless of the orientation of the orthogonal polarization axes, so as to conceptually divide the area of ​​the footprint of the light output from the non-uniform / patterned retarder 114 that forms the optical signal at a sensor (e.g., sensor 20) into a one-dimensional array of two equal-width rows extending parallel to the Y dimension. If necessary, the number of rows by which the area is divided, and which must comply with the above requirement, can be increased (e.g., to four or eight) to ensure a more uniform distribution of the different polarization directions.

[0141] In summary, by generating a uniform mix of polarizations for any input polarization state, the DOP of light leaving the nonuniform / patterned retarder can be minimized. This can be achieved by varying the fast axis direction of the nonuniform / patterned retarder across the field of view in a way that achieves this. One way to achieve this is a random distribution of fast axis directions. If the fast axis direction selection were truly random, it could result in a very effective nonuniform / patterned retarder, but it could also result in a nonuniform / patterned retarder that is less effective. Therefore, for consistency and predictability, it may be preferable for the variation of fast axis direction across the nonuniform / patterned retarder to be non-random. The ideal single-pass solution would be one that mixes the fast axes as evenly as possible.

[0142] Therefore, instead of randomly distributing the fast axis directions, pixel fast axis directions can be evenly distributed with a constant angular difference between pixels. This can have the advantage of evenly distributing the polarization angles when light reaches the scale. For example, the distribution can align pixels at angles of 0°, 1°, 2°, 3°, etc., with a 1-degree spacing between pixels. Alternatively, this is possible with a finer sampling size in fractions of degrees, or even larger increments between actual fast axis angles. For example, larger increments of 75° between pixels would result in fast axis values ​​of 0°, 75°, 150°, 225°, 300°, etc. However, this is not the optimal way to express these values. The effective fast axis of a waveplate is modulo 180°. That is, a waveplate will behave the same as long as its phase is rotated by an integer multiple of 180°. Therefore, the previous sequence is more appropriately combined as the remainder after dividing 0°, 75°, 150°, 45°, and 120° by 180°. The fast axis values ​​generated in this manner will be described herein as a phase-wrapping linear function. To provide a good mix of polarization states, it may be preferable for the fast axis values ​​of the non-uniform / patterned retarder not to repeat over its length (e.g., in the Y dimension in the described embodiment). In other words, it may be preferable for the fast axis values ​​of the non-uniform / patterned retarder to vary according to a non-repeating design / form / pattern. Therefore, preferred embodiments of non-uniform / patterned retarders include fast axis values ​​that vary according to a phase-wrapping linear function that does not provide repetition over at least the length of the region of the light footprint output from the non-uniform / patterned retarder that forms the optical signal at the sensor. As will be appreciated, the 75 incremental values ​​would only provide 12 different values ​​of the fast axis before the repetition begins, and thus would only provide 12 pixels of non-uniform / patterned retarders without repetition of fast axis values. Therefore, if a non-uniform / patterned retarder with more pixels is desired before the fast axis value starts to repeat, different fast axis increment values ​​between pixels are required and can be selected according to user requirements.For example, a fast axis increment value of 42 degrees would provide 30 distinct fast axis values ​​before any repetition occurs (thus providing a 30 pixel non-uniform / patterned retarder without any repetition of fast axis values).

[0143] Optionally, in an alternative embodiment, a machine learning optimization process can be used to assist in the design of the optical retarder element (also referred to as a non-uniform / patterned retarder). Accurately modeling diffraction effects can be beneficial to create a design optimized for a specific system geometry. At the heart of the optimization loop in one embodiment is a one-dimensional physical optics simulation based on the target system geometry. The input is a fully polarized plane wave simulating collimated light incident on the non-uniform / patterned retarder. The interaction of this plane wave with the non-uniform / patterned retarder is simulated using Jones calculus. A field is then propagated using non-paraxial Fresnel propagation or to a plane where the scale is located in the target encoder geometry. The field at this plane is recorded for evaluation. The field is then propagated to a second non-uniform / patterned retarder that takes the same form as the first retarder but has a position offset equal to the target offset in the encoder system. The interaction with the field is simulated before propagating to the plane of the lens where the field is again recorded.

[0144] This entire process is repeated as the polarization of the input plane wave is moved through a 180-degree range of possible input polarizations. This data can be used to estimate the DOP at the scale and detector as a function of length scale. The DOP at the scale and detector form the two metrics for optimization, and the length scale at which the DOP is evaluated depends on the sensor size and scale details. A weighted sum of these two values ​​forms the minimized target. Optionally, the DOP at the scale is weighted more heavily than the DOP at the detector.

[0145] A simulated annealing optimization approach can be used. At each iteration, the fast axis direction of the pixel can be changed to a new random value, and a target value can be calculated and compared to the previous target value. If the new value is smaller than the previous value, the new solution is accepted. If the new value is greater than (or equal to) the previous value, the solution may or may not be selected based on chance. The probability of acceptance is given by:

[0146]

number

[0147] where Δ = new target value - old target value, and T = current temperature (or "free parameter"). The probability of acceptance depends on how much worse (higher) the new target value is than the old target value and the current temperature. The temperature starts high and is lowered after each cycle. The initial temperature is set high to allow the optimization to freely explore the solution space before lowering the temperature forces the optimization to converge to a solution. This approach reduces the probability of converging to a local minimum instead of a global minimum. In a specific implementation, the temperature is lowered after a random trial is performed on every pixel in the retarder. This completes a single loop. The process is then repeated many times until the algorithm converges to a solution. The output is an aperiodic solution that generally has strong performance characteristics but cannot be expressed analytically.

[0148] In the above-described embodiments, the waveplate pixels 19, 119 include half-waveplates. However, it will be understood that the waveplate pixels 19, 119 can instead include quarter-waveplates, or indeed any fractional waveplate. Furthermore, in the above-described embodiments, the waveplate pixels 19, 119 all have the same waveplate fraction (i.e., the same retardance value), with the fast axis being varied from waveplate pixel to waveplate pixel. However, this need not necessarily be the case. Instead, the fast axis of each pixel and the waveplate fraction (or "retardance") of the waveplate pixel can vary. Optionally, in another embodiment, the non-uniform / patterned retarder can be configured such that the fast axis and waveplate fraction can vary from waveplate pixel to waveplate pixel. The waveplate pixels can include birefringent materials such as quartz, mica, liquid crystal, etc. Optionally, the waveplate pixels can include structures configured to provide birefringence-like properties, such as metamaterials. For example, it is known that waveplates can be provided using nanocasting lithography.

[0149] It will be appreciated that other types of sensors can be used instead of the electrical gratings described above, for example in embodiments where a modulated spot is created by the readhead optics instead of an interference fringe pattern, a bulk sensor photodiode can be used to detect the intensity of the modulated spot.

[0150] In the described embodiment, light source 12 emits electromagnetic radiation (EMR) in the near-infrared range. However, it will be appreciated that this need not be the case and that EMR can be emitted in other ranges, for example, any range from infrared to ultraviolet. It will be appreciated that the selection of an appropriate wavelength for light source 12 can depend on many factors, including the availability of suitable gratings and detectors operating at EMR wavelengths.

[0151] In the above described embodiments, the scale is a reflective scale. However, this does not have to be the case. For example, the scale could be a transmissive scale, with the light source and sensor located on opposite sides / faces of the scale.

[0152] In the above-described embodiments, the encoder includes an incremental scale with one or more reference marks (or, optionally, without). However, the non-uniform / patterned retarder may be used with other types of optical encoders, for example, absolute encoders that use a VCSEL instead of an unpolarized light source such as an LED. See, for example, the absolute encoders described in U.S. Pat. No. 5,279,044 or U.S. Pat. No. 1,013,257. If a VCSEL light source is used, the use of the non-uniform / patterned retarder of the present invention may be beneficial. In particular, the use of a non-uniform / patterned retarder can avoid any adverse effects caused by polarized light impinging on the scale. Furthermore, optionally, the non-uniform / patterned retarder may be particularly useful in absolute encoders when encoding one or more of the diffraction orders from the scale with a polarization state (e.g., for subsequent filtering purposes), such as in the embodiments of FIGS. 7 and 11 above.

[0153] The above-described embodiments use a non-uniform / patterned retarder as a polarization manipulator to reduce the degree of polarization. Instead of a non-uniform / patterned retarder, other types of optical elements, such as a diffuser or a time-based polarization manipulator, can be used. As will be appreciated, for a time-based polarization manipulator, the modulation rate will need to be faster than the acquisition bandwidth of the encoder's sensor. Typically, a rate of at least 1 MHz should be sufficient. Fiber-coupled time-based polarization frequency band converters, such as those available from FIBREPRO Inc., exist that are suitable for fiber-optic position encoders. Because time-based manipulators require electronics that are not suitable for compact encoders, and because, in some embodiments, diffusers can provide significant adverse ray deflections that adversely affect the formation of interference fringe patterns, spatial polarization manipulators, particularly non-uniform / patterned retarders, have been found to be preferred.

[0154] The embodiments of Figures 7 and 11 described above use a refractive lens 109 to relay the diffracted orders towards the sensor. However, it will be appreciated that this is not necessarily the case. For example, Figure 18 shows a schematic of the optical system of an encoder device including an incremental scale 104 and a readhead 206' according to another embodiment of the present invention. This embodiment is similar to Figure 11 (like parts share like reference numerals) except that it does not rely on a refractive lens 109 to relay the diffracted orders towards the sensor. In this embodiment, the zeroth and ±1st diffraction orders are generated from light from a light source that strikes and is reflected by the scale 104. As shown, the zeroth and ±1st diffraction orders are reflected by a first plane pl 1 The diffraction order encoder 216 is located in a first plane pl 1According to the embodiment of FIG. 11, the diffraction order encoder includes a polarizing filter 216a (e.g., with a "vertical" polarization axis). The polarizing filter 216a is matched to the zeroth diffraction order so as to encode the zeroth diffraction order in a first (e.g., vertical) polarization direction. The diffraction order encoder further includes second and third polarizing filters 216b, 216c matched to the +1st and -1st diffraction orders. The second and third polarizing filters 216b, 216c are configured with the same orientation as each other so that the ±1st diffraction orders have the same (e.g., "horizontal") polarization direction as each other but different from the polarization direction of the zeroth diffraction order. The diffraction grating G2 is positioned to focus the polarized ±1st diffraction orders. The zeroth diffraction order is not polarized. Therefore, the zeroth and ±1st diffraction orders are polarized relative to the plane pl in which the incremental sensor 20 and the reference mark sensor 122 are located. 2 11 embodiment, a first polarizer 120 may be positioned in front of the incremental sensor 20 and configured to filter the 0th diffraction order from reaching the incremental sensor 20, and / or a second polarizer 223 may be positioned in front of the reference mark sensor 122 and configured to filter the ±1st diffraction orders from reaching the reference mark sensor 122. For clarity of illustration, the light source 12 and non-uniform / patterned retarder have been omitted from FIG. 18. As will be appreciated, a separate refractive lens or grating may be used to collimate the light from the light source towards the scale, if desired.

[0155] The embodiments described thus far have related to incremental encoders that include an incremental scale with (optionally without) one or more reference marks. The present invention can also be used with absolute encoders that include absolute scales, such as those shown in FIGS. 19 and 20. In the embodiment of FIG. 19, the absolute scale 504 includes conceptually periodically arranged features, with selected features removed to encode unique / absolute position data along the measuring length of the scale. The data can be in the form of, for example, a pseudo-random sequence or discrete code words. Details of such scales are described in U.S. Pat. Nos. 7,499,827 and 5,279,044. The readhead 506 (the body of which has been omitted for clarity) shares several parts identical to those described above in connection with other embodiments of the present invention, and therefore, similar parts share the same reference numerals. For example, the readhead 506 includes a VCSEL light source 12, a lens 109, an optical retarder element (also referred to as a non-uniform / patterned retarder) 214, a diffraction order encoder 216, an incremental detector 20, and a primary sensor filter 120, which are configured and arranged in the same manner as described above in connection with the embodiment of FIG. 11. Thus, the zeroth order term is blocked from contributing to the formation of the signal detected by the incremental sensor 20. As with the other embodiments, the primary sensor filter 120 includes a polarizing filter having a polarization axis perpendicular to the polarization axis of the first polarizing filter 216a of the diffraction order encoder 216. Notably, in this embodiment, the first polarizing filter 216a has a vertical polarization axis, and the primary sensor filter 120 has a horizontal polarization axis.

[0156] In this embodiment, the readhead 506 further includes an absolute sensor 520. The absolute sensor 520 includes a photodiode array (in this embodiment, a one-dimensional photodiode array, but could also be two-dimensional) that is impinged by an image (or "pseudo-image," see above). As is known and described, for example, in U.S. Pat. Nos. 7,499,827, 5,279,044, and 10,989,567, the image of the scale can be processed to extract an absolute / unique code and thereby determine the absolute position. The absolute position can be combined with the incremental position determined from the incremental detector 20 to provide a fine-pitch absolute position. Optionally, once the absolute position is determined, subsequent positions can be determined independently by monitoring the output from the incremental detector 20.

[0157] As with the reference mark embodiment described above, it may be beneficial to allow the zeroth order term to contribute to the signal impinging on the absolute sensor. Additionally, as with the reference mark embodiment described above, it may be beneficial to prevent the ±1st diffraction orders from contributing to the signal impinging on the absolute sensor 520. Accordingly, a secondary sensor filter 523 may be provided in front of the absolute sensor 520, having a polarization axis orthogonal to the polarization axes of the second and third polarization filters 216b, 216c of the diffraction order encoder 216. Notably, in this embodiment, the second and third polarization filters 216b, 216c have horizontal polarization axes, and the secondary sensor filter 523 has a vertical polarization axis.

[0158] FIG. 20 illustrates another embodiment of an absolute encoder 600 including an absolute scale 604. The absolute scale 604 is substantially identical to that described above in connection with FIG. 21, except that in this embodiment, the scale includes a multi-track arrangement, with absolute code features provided in an absolute scale track 603 provided on either side of a pure incremental track 605 (like parts share the same reference numerals). As shown, the absolute scale track 603 also includes finer pitch / period incremental features in spaces between the coarser pitch / period features of the absolute scale track 603. By way of example, the coarse features of the absolute scale track that encode absolute position information may have a nominal period of approximately 32 μm, and the finer pitch / period incremental features may have a period of approximately 8 μm. As a result, two sets of diffraction orders are effectively generated: a first set of diffraction orders generated by a first series of position features (e.g., 8 μm incremental scale features) and a second set of diffraction orders generated by a second series of position features (e.g., absolute scale features). In fact, in this embodiment, the zeroth diffraction order of the first set of diffraction orders passes through the first polarizing filter 216a of the diffraction order encoder 216, the ±1st diffraction orders of the first set of diffraction orders pass through the second and third polarizing filters 216b, 216c, while the zeroth diffraction order and the ±1st diffraction orders of the second set of diffraction orders all pass through the first polarizing filter 216a of the diffraction order encoder 216. The higher diffraction orders of the first set of diffraction orders (e.g., the ±3rd diffraction orders, the ±5th diffraction orders) are blocked by the opaque substrate 115. The ±3rd diffraction orders of the second set of diffraction orders pass through the second and third polarizing filters 216b, 216c of the diffraction order encoder 216. Although the diffraction orders higher than the ±3rd diffraction orders of the second set of diffraction orders have substantially insignificant power, what is present will be blocked by the opaque substrate 115. As will be appreciated, which diffraction orders pass through which filters is system dependent and may depend on a variety of factors, including the size and / or position of the polarizing filters.

[0159] As shown, the sensor arrangement of the embodiment of FIG. 20 differs slightly from the embodiment of FIG. 19 in that the readhead of the absolute sensor 620 is split in two with the incremental sensor 22 located between them (similar to the embodiments of FIGS. 7 and 11 ). A variation of the embodiment of FIG. 20 is for the absolute scale features to extend completely across the incremental scale features (i.e., so that there is no pure incremental scale track 605 between the two absolute tracks 603). Rather, as in the embodiment of FIG. 19 , the absolute scale features can be completely embedded throughout the incremental features. In this case, the comments above regarding the two sets of diffraction orders still apply because, in this embodiment, the incremental and absolute position features have different nominal periods. If desired, an alternative configuration could arrange the incremental sensor 18 and absolute sensor 620 according to the embodiment of FIG. 19 (i.e., with two sensors arranged side-by-side).

[0160] The above-described embodiments utilize a lensless or single-lens system through which light from the light source passes on its way to and from the scale. It will be appreciated that other optical configurations are possible, such as the one shown schematically in FIG. 21 (which illustrates an absolute encoder device including an absolute scale). The encoder 300 of FIG. 21 includes an absolute scale 304 that includes conceptually periodically arranged features, with selected features removed to encode unique / absolute position data along the measuring length of the scale. The data may be in the form of, for example, a pseudo-random sequence or discrete codewords. Details of such scales are described in more detail in U.S. Pat. Nos. 7,499,827 and 5,279,044. The readhead 306 (whose body has been omitted for clarity) shares several parts identical to those described above in connection with other embodiments of the present invention, and therefore, similar parts share the same reference numerals. For example, the readhead 306 includes a polarized VCSEL light source 12 and a non-uniform / patterned retarder 14. A first lens 309 is provided to collimate the light from the VCSEL 12. An optical beam splitting section 307 is provided which is able to pass light from the VCSEL 12 through to illuminate the scale 304 and redirect light reflected by the scale to a sensor 320 (in this case a one-dimensional array of photodiodes, e.g. a complimentary metal-oxide-semiconductor (CMOS) sensor). The sensor 320 is configured to sense an image of the scale formed in situ by a second lens 310. The image is processed by a processor device to determine the relative positions of the readhead 306 and the scale 304 in a known manner (e.g. as described in U.S. Pat. No. 1,098,567). As shown, this configuration is less compact than other embodiments, but has the advantage that the non-uniform / patterned retarder can be placed after the collimating lens, avoiding a double pass of the non-uniform / patterned retarder.

[0161] The present invention has been described above in relation to a linear encoder device, however, the present invention is equally applicable to rotary encoder devices including both ring encoders (which form scale features on the outer cylindrical surface of a ring member) and disk encoders (which form scale features on the planar surface of a disk member).

Claims

1. a scale readable by the readhead and including a series of position features extending along the measurement dimension; The read head and Equipped with The read head i) a polarized light source that emits polarized light toward the scale; ii) a sensor configured to detect optical signals from positional features of the scale that can be used to determine the relative position of the scale and the readhead along the measurement dimension; and iii) a polarization manipulator located in an optical path between the polarized light source and the sensor, configured such that the degree of polarization of light output from the polarization manipulator is less than the degree of polarization of the polarized light received by the polarization manipulator, regardless of the predominant input polarization direction of the footprint of the polarized light received by the polarization manipulator; Including, Position measuring encoder device.

2. the polarization manipulator is configured such that a degree of polarization of light output from at least a region of the optical footprint output from the polarization manipulator forming the optical signal at the sensor is less than a degree of polarization of the polarized light emitted from the polarized light source.

2. The position measuring encoder device of claim 1.

3. light from the polarized light source is configured to pass through the polarization manipulator a first time on the way to the scale and pass through the polarization manipulator a second time after being reflected from the scale; the polarization manipulator is configured such that after the light passes through the polarization manipulator twice, the degree of polarization of the light output from at least a region of the footprint of the light output from the polarization manipulator forming the light signal at the sensor is less than the degree of polarization of the polarized light emitted from the polarized light source.

3. A position measuring encoder device according to claim 1 or 2.

4. configured to conceptually divide an area of ​​the footprint of light output from the polarization manipulator forming the optical signal at the sensor into a one-dimensional array of n columns of equal width, wherein the degree of polarization of the light output from each of the n columns is less than the degree of polarization of the polarized light emitted from the polarized light source, n being 2, and wherein the one-dimensional array extends perpendicular to the measuring dimension of the scale; A position measuring encoder device according to any one of claims 1 to 3.

5. The requirements of claim 4 apply for both the first and second passes of the polarization manipulator.

5. A position measuring encoder device according to claim 3 or 4.

6. n is 4, and optionally, n is 8; 6. A position measuring encoder device according to claim 4 or 5.

7. in a dimension perpendicular to said measuring dimension, in the area on the sensor illuminated by light from the polarization manipulator, and - configured such that, without a polarizer in the optical path between the polarization manipulator and the sensor, the polarization direction of the light from the polarization manipulator varies over a length scale that is at least five times smaller than the range of the sensor; A position measuring encoder device according to any one of claims 1 to 6.

8. the polarization manipulator includes an optical retarder element; A position measuring encoder device according to any one of claims 1 to 7.

9. along at least one dimension, the optical retarder element comprises a non-uniform fast axis; 9. A position measuring encoder device according to claim 8.

10. the at least one dimension along which the fast axis varies is perpendicular to the measurement dimension; 10. A position measuring encoder device according to claim 9.

11. the fast axis varies along only one dimension; 11. A position measuring encoder device according to claim 9 or 10.

12. The fast axis varies according to a non-repeated design. A position measuring encoder arrangement according to any one of claims 9 to 11.

13. the fast axis varies according to a phase wrapping linear function of position along at least one dimension of the optical retarder element; 13. A position measuring encoder device according to claim 12.

14. the optical retarder element includes a one-dimensional series of discrete corrugations, the series of discrete corrugations extending in a dimension perpendicular to the measuring dimension; A position measuring encoder arrangement according to any one of claims 8 to 14.

15. the readhead further includes a lens, the lens being positioned between the polarized light source and the polarization manipulator; A position measuring encoder arrangement according to any preceding claim.

16. the polarization manipulator is positioned in the optical path of the light to and from the scale; A position measuring encoder arrangement according to any preceding claim.

17. the polarized light source is a laser light source, optionally a Vertical Cavity Surface-Emitting Laser (VCSEL); 17. A position measuring encoder device according to any preceding claim.

18. a scale including a series of location features extending across the measurement dimension; The read head and Equipped with The read head i. a polarized light source that emits polarized light toward the scale; ii. A sensor configured to detect an optical signal from a position feature of the scale; iii. an optical retarder element located in an optical path between the polarized light source and the sensor, the optical retarder element having a non-uniform fast axis and / or retardation characteristic along at least one dimension, configured such that light output from the optical retarder element contains a mix of polarization directions regardless of the predominant input polarization direction of the footprint of polarized light impinging on the optical retarder element; Including, Position measuring encoder device.

Images