NO136061B - - Google Patents

Description

The present invention relates to a position measuring transducer which has relatively movable winding elements, where one of said elements includes a continuous reference winding, where a second of said elements consists of a first pair of polyphase winding sections and a second pair of polyphase winding sections, where a winding section in both pairs is in circuit with each other, and where the second winding section in both pairs is in circuit with each other, where said first and second pairs of winding sections are superimposed with respect to each other in separate layers.

In practice, position-measuring transformers on the single-phase element have comprised a single winding made up of evenly spaced series-connected active conductors, where adjacent conductors lead in opposite directions. For linear devices, in practice the single-phase element is called the "scale" and for rotating devices it is called the "rotor".

The other relatively mobile element, called the multiphase element,

in position-measuring transformers usually comprise two multi-phase windings, each phase-shifted in space in relation to the other, whereby two different space phases are presented to the second term. Conventionally, the polyphase element is called the "slider" in linear devices and the "stator" in rotary devices.

The phase shift between the multi-phase windings can be a quarter of a space cycle for the single-phase winding. When the polyphase windings are shifted by a quarter cycle, they are conventionally referred to as sine and cosine windings. Although the sine and cosine windings are conventional,

can phase shifts on e.g. 120° be possible.

When a winding on one part of a position-measuring transformer is magnetized with a primary alternating voltage signal, a switching signal, also called a coupling signal, is induced in each winding on the other part of the position-measuring transformer, in whose immediate vicinity the winding is located. For accurate measurements, it is desirable that the coupling between the windings varies exactly as a non-premagnetized sine curve as a function of the relative space displacement between the windings over each space cycle. The space cycle is. equal to twice the space P between the conductors, i.e. the space cycle is equal to 2P. The space frequency, or more precisely the space fundamental frequency, is defined as the inverse, 1/2P, of the spacing period. For an ideal system, the coupling between the windings is a perfect sine curve with a space fundamental frequency of 1/2P.

It is well known from the past that position measuring transducers tend to have a coupling which is not precisely sinusoidal or which is not zero biased. Usually, the coupling comprises coupling components originating from higher order harmonics of the space fundamental frequency, in particular higher order different harmonics. Unwanted coupling components also originate from coupling at a lower frequency than the fundamental frequency, in particular from a constant coupling (zero frequency coupling). Constant coupling is the coupling that originates from a constant field that does not vary as a function of the spatial position (hence zero frequency) or from a variable field with a constant premagnetization component (zero frequency component). Constant coupling is sometimes referred to as entron's sloyfe coupling. Couplings other than the space fundamental frequency between the windings in a position-measuring transformer cause unwanted errors in position measurement and must therefore be avoided.

US patent no. 2 650 352 deals with a transducer with a continuous winding on one element connected inductively to a continuous winding on the other element. In transducers of this type, the measurement accuracy is limited because the coupling between the two windings as a function of their relative spatial positions is not precisely sinusoidal. The lack of sinusoidal coupling is partly due to the inductive constant coupling component (zero frequency component) which occurs between the windings. Such inductive constant coupling causes an error component which is defined as a once-per-cycle or basic error component.

In US patent, no. 2,799,835, several techniques are discussed for achieving a more accurate sinusoidal connection between the links in a position-measuring transformer. To avoid constant coupling, one or more of the windings is divided into several winding sections, where half of the sections in each winding are connected positively in relation to the constant coupling, and the other half are connected negatively in relation to the constant coupling. When the positive and negative sections are electrically connected, the constant coupling components tend to cancel each other out.

US patent no. 2,799,835 also deals with further techniques for making the coupling between the windings in position-measuring transducers more sinusoidal, particularly with regard to unwanted coupling at higher-order harmonics of the fundamental frequency. The distance ratio between conductor and space, the slope of the active conductors and the distance between electrically connected groups of winding sections are examples of applied technique.

US Patent No. 2,915,721 relates to a transducer in which half-current feedback conductors are used to establish field monsters which tend to reduce the constant coupling between the elements of a position-measuring transformer. The half-current return conductors are arranged in parallel with the end conductors of the single-phase winding in a position-measuring transformer.

In US patent no. 2 915 722 neutralization of the constant coupling is achieved by half of the winding sections for each multiphase winding (sine and cosine) being connected in opposition to the other half with respect to the constant coupling.

In addition to the purpose of achieving precise sinusoidal coupling between the windings, two-phase systems also preferably have two windings in precise quadrature, i.e. they are separated by exactly a quarter of a space cycle. Lack of quadrature between multi-phase windings produces measurement errors which are termed phase-shifted second-harmonic errors.

To ensure more accurate quadrature between the windings, the aforementioned US patent no. 2,915,722 uses a kind of quadrature compensation. The lack of accurate quadrature in this patent stems from variations in length, such as variations caused by temperature change. In a similar way, US Patent No. 3,441,888 uses a kind of quadrature compensation in a transducer provided with a large number of sine and cosine winding sections arranged opposite each other.

Although all the above-described technology has contributed to position-measuring transformers that can perform precise measurements,

is it desirable to achieve further improvement in accuracy and manufacturing technique.

A problem with previous transducers is their susceptibility to error-creating anomalies, which give a lack of uniformity, and which are characterized by irregularities in the measurement accuracy from one spatial position to another. Anomalies occur e.g. at the connection between the end-to-end adjacent rod scales that make up the continuous reference winding of a transformer. These connections tend to produce anomalies in the connection field. Other types of anomalies occur e.g. of defects or irregularities in materials, of deformity in conductors, in unevenness in the air gap, which e.g. can happen due to a break in a tape scale or due to terminal pins and wires.

To avoid previous errors, the unwanted coupling in one winding section in one room area is compensated by the unwanted coupling in another winding section in a different room area. The degree of compensation and consequently the degree of error reduction depends on a non-varying coupling ratio for the two different areas. However, a number of factors cause variations in the couplings in different regions and therefore produce detrimental effects on the operation of transducers which are based on compensation of one different region against another. E.g. will an anomaly that only connects to one of the two spatial areas at a time disturb the compensation relationship.

Previously manufactured transducers capable of accurate measurement have generally been of the non-continuous type, ie with windings formed by a large number of winding sections. Non-continuous windings are a problem because the field monsters of the first and last active conductors in each winding section are not in opposition and are therefore irregular compared to the precise alternating monsters in the intex active conductors. These irregularities accompanying the first and last conductors of each winding section produce measurement errors, especially when coupled with other anomalies, such as the connection between two scales on a reference winding. Because of these errors, the greater the number of winding sections, the greater the number of errors due to irregularity in the end effect of the first and last conductor.

In addition to the end effect problem, previously manufactured transducers have used sine and cosine winding sections that are more spaced than the desired minimum of one quarter space cycle. When these sine and cosine winding sections transform over an anomaly, the sine section is disturbed at a different time than the cosine section, causing an error due to the unwanted change in the ratio between the sine and cosine coupling.

In light of the above-mentioned background to the invention, it is an object of this invention to provide transducers which, in various combinations, have reduced sensitivity to anomalies, smoother accuracy, quadrature (or other phase) compensation, constant cobbling reduction and cancellation of unwanted harmonics.

The transducer according to the invention is characterized primarily by the fact that each of the winding sections in each of said pairs comprises a continuous arrangement of evenly spaced conductor arms, that said first pair of multiphase winding sections are placed in approximate spatial quadrature with respect to each other, that the second pair of multiphase winding sections is placed in the same approximate spatial quadrature with respect to each other, that a first winding section in the first pair of multiphase winding sections is arranged relative to a first winding section in the second pair of multiphase winding sections with an offset a and that the second winding section for said first pair of multiphase winding sections is arranged relative to the second winding section in said second pair of multiple winding sections with the same displacement a, where oc is the same in the above cases and represents 90° plus an error which is the same in both cases.

According to further features of the invention, the transducer is characterized by the fact that said first winding sections in said multiphase windings are arranged closest to said reference winding and that the distance ratio between conductor and space for said first winding sections is greater than the active distance ratio between conductor and space for said second winding sections, that the average of the distance conditions between conductor and space for said first and second winding sections for each of said multiphase windings is approximately equal to 2:1, whereby coupling of the third harmonic between said multiphase windings and said reference winding tends to be neutralized, that said winding sections have the form of a spiral arrangement, that a conductor arm in each of all said winding six ions is placed in the same half-cycle space and that successive space cycles of said multiphase windings each contain a conductor arm from each of said winding sections.

Further features of the present invention will be apparent from the following description with reference to the drawings.

Fig. 1 shows, according to the invention, a schematic representation of a single winding (completely drawn up) for localization on one element inductively connected to another winding for localization on another element, where the latter winding is made up of a first winding section (dotted) on one layer arranged opposite another winding six ion (broken) on another layer.

'Fig. 2 shows a representation of the alternating field monster generated by the transformer in fig. 1.

Fig. 3 shows waveforms which are representative of the switching signals in the one in fig. 1 shown apparatus. Fig. 4 schematically shows two rows of approx. 90° phase-shifted active conductors arranged on a plate in appropriate proportions to provide, after further steps, two layers of conductors for a multi-layer transformer element according to the invention. Fig. 5 schematically shows another monster identical to the monster in fig. 4, but shifted from left to right on the plate to facilitate the understanding of fig. 6. Fig. 6 schematically shows a superimposed combination of the monsters in fig. 4 and 5 with the addition of end conductors to form four separate ones

winding sections on two layers.

Fig. 7 schematically shows the windings for a transformer element, where the windings are arranged in a four-layer structure constructed from the monsters in fig. 4, 5 and 6. Fig. 8 shows a vector diagram describing the quadrature compensation feature present at the windings in fig. 7. Fig. 9 shows another vector diagram to explain the quadrature compensation feature of the windings in fig. 7. Fig. 10 shows, seen from above, the four layers which are combined in a superimposed relationship to form the one in fig. 7 schematically shown structure. Fig. 11 shows, seen from above, overlying terminal connections suitable for use with the winding sections in fig. 10. Fig. 12 shows a representation of an enlarged cross-section of the active conductors in a reference winding over four pairs of active conductors, one pair from each of the four winding sections in fig. 10, where the four layers are arranged in one four-layer structure as shown in fig. 7. Fig. 13 shows a perspective view seen from above and from the front of a cropped transducer element comprising four layers of winding sections of the one in fig. 10 shown type. Fig. 14 shows a perspective view seen from the end and from the front of that in fig. 13 showed a winding element facing another winding element. Fig. 15 shows a schematic representation of a monster on a layer with two rows of radial active conductors that can be used to construct a multipole rotating transducer. Fig. 16 shows another schematic monster for rotary transducers substantially identical to the monster in fig. 15. Fig. 17 shows a schematic representation of the superimposed monsters in fig. 15 and 16. Fig. 18 shows a simplified representation of superimposed monsters that appear by folding the two in fig. 17 showed radial monsters facing each other, both along a line 41 (shown in Fig. 20). Fig. 19 schematically shows four layers of winding sections formed by adding end conductor parts to the rows in fig. 17, where the layers are separated for clarity in the same order as in fig. 10. Fig. 20 shows a schematic representation of the four winding sections in fig. 19 arranged against each other around a common center to form a four-layer structure of a rotating position transducer. Fig. 21 shows a vector representation of the quadrature compensation feature in the one in fig. 20 shown device. Fig. 22 shows a schematic representation of a reference winding under a two-layer winding section element, where two monsters as in fig. 6 are placed side by side. Fig. 23 shows an error curve for a transducer according to the invention of the one in fig. 7 showed type as well as an error curve for a typical previous measuring converter and shown above the transducer measuring device. Fig. 24 shows a schematic representation of a winding element used as one layer in a multi-layer rotary transducer according to the invention with one cycle per revolution. Fig. 25 schematically shows a sample of the one in fig. 24 shown type, but rotated 90°. Fig. 26 schematically shows a holder and two double layers of laminates which can be used to produce a four-layer converter according to the invention with windings of the one in fig. 24 and 25 shown type. Fig. 27 shows the holder in fig. 26 with the laminates arranged for the second step in the manufacturing process which produces quadrature-<k>om<p>ensation according to the invention. Fig. 28 schematically shows the four-stage configuration that appears after the processing steps in fig. 26 and 27. Fig. 29 schematically shows the way in which the multi-layer construction in fig. 28 are combined to form a four-layer structure. Fig. 1 shows a representation of a position-measuring transformer according to the invention, in which a first winding 2 is arranged to magnetically connect to another winding 6, the windings 2 and 6 both sitting on relatively movable (not shown) elements. It

second winding 6 comprises a first winding section 8 (shown dashed) and a second winding section 9 (shown dotted). Winding sections 8 and 9 are both typical printed copper conductors in different layers, the layers sitting on opposite sides of an insulating layer to form a laminate. The relevant details of typical layer thicknesses are discussed below in connection with fig. 13 and 14.

First winding 2 is made up of active conductor parts with a prefix

3- and denoted 3-1, 3-2, ..... 3-8. The active leadership parties 3-

is connected in series using end conductor sections with prefix 4- and as in fig. 1 is denoted 4-1, 4-2, ..... 4-7. The end conductors 4-

are arranged alternately along the edges defined by opposite ends of the active conductor parts 3-, so that adjacent active conductor parts 3- lead in opposite directions. Adjacent active conductor parts therefore denote opposite poles, and the distance between adjacent conductor parts on either side of a given active conductor part is equal to a full space cycle.

In a similar way as for the first winding 2, each of the winding sections 8 and 9 for the second winding 6 is also made of active conductor parts connected by means of end conductor parts along alternating edges.

In particular, the winding section 8 is provided with active conductor parts with the prefix 18- and designated 18-1, 18-2, ..... 18-6, while the winding section 9 is provided with active conductor parts with the prefix 19- and designated 19-1, 19 -2, ..... 19-6.

The active conductor parts 18-1 on winding 8 are connected in series at

using end conductor sections with prefix 16-, and the active conductor sections on winding section 9 are connected in series using end conductor sections with prefix 17-. Final leader parties 16-1 to 16-6

and 17-1 to 17-6 are shown in fig. 1.

In actual practice, one of the windings 6 or 2 is as much longer than the other as the distance over which measurement or movement is desired.

According to the invention, the first and second winding sections 8 and 9 are arranged so that the active conductor parts 18-2 and 19-1 e.g. mainly have the same distance in relation to the active leading parties

3- on first winding 2. Furthermore, the winding sections 8 and 9 are connected in series at the terminals 11, so that the active conductor parts 18-2 and 19-1 e.g. both lead in the same direction. Similarly, both the other pairs of active leader parties 18-3 and 19-2, 18-4 and 19-3, 18-5 and 19-4 and 18-6 and 19-5 have leaders in the pair leading in the same direction. In this way, each conductor in the pair has essentially the same connection relationship to the active conductor parts in the first winding 2.

The accuracy with which conductors 19-1 and 19-2 are superimposed in the same space is not critical, but the further they are spaced apart, the smaller the combined coupling resulting from the pair.

While the active conductor parts of the winding sections 8 and 9 in fig. 1 are generally connected and spatially arranged to be additive in relation to each other, the end conductor parts 16- and 17- are arranged so that they lead in opposite directions. More particularly, the direction of the end conductor portions 16-1 is opposite to the direction of the end conductor portions 17-2. Similarly, the directions of the pairs 17-1 and 16-2, 16-3 and 17-4 etc. are all opposite. The effect of having the active conductor parts arranged additively, while the end conductor parts have alternating opposite directions, will appear from fig. 2.

When in fig. 1 an alternating voltage signal is applied between the terminals 21 and 22, .occurs in the active conductor parts 18- and 19-, as shown in fig. 2, field vectors 18' and 19' which change direction every half-period, and cause a voltage on the terminals 2 3 and 24 in fig. 1. The fields originating from the end conductor parts also change direction every half period, so that the net sum of the end conductor parts' 16' and 17' fields in fig. 2 considered in a clockwise or counterclockwise direction is close to zero.

The field monster in fig. 2, which is produced by two superimposed six windings 8 and 9 which comprise the second winding 6 in fig. 1, is similar to the field sample generated according to US Patent No. 2,915,721, where this patent uses the different technique of half-strbm return conductors.

The production of the field sample in fig. 2 can be used to reduce or neutralize the effects of constant coupling between the windings in a position-measuring transformer. The reduction in unwanted constant coupling can e.g. is explained in connection with fig. 3. 1 fig. 3, the waveform 25 represents the amplitude of the coupling wave that occurs between the active conductor parts 3- in first winding 2 and the active conductor parts 18- and 19- in the second winding 6. When second winding 6 and first winding 2 are moved in relation to each other so that the active conductor part 3-2 is precisely superimposed on the active conductor part 18-1, the windings are defined here in the description for simplicity to be at the O-degree point where maximum coupling occurs. The maximum coupling between windings 2 and 6 is indicated by the positive peak in curve 25 in fig. 3 at the O point. When the windings in fig. 1 are moved relative to each other so that the active conductor portion 3-2 is halfway between the active conductors 18-1 and 18-2, the windings are defined to be at the 90-degree point and have zero coupling as shown in fig. 3. When the active conductor portion 3-2 is moved to the 180-degree point above the active conductors 18-2 and 19-1, a maximum negative peak occurs as shown in fig. 3. Zero connection occurs again in the 270-degree point, e.g. when the active conductor 3-2 is halfway between the active conductors 18-2 and 18-3. Finally, the coupling bolt 25 again reaches a positive maximum at the 360-degree point, e.g. when the active leader party 3-2 is aligned with the active leader parties 18-3 and 19-2.

The transducer in fig. 1 can e.g. is used to define zero positions with an equal distance. Zero positions with the same distance have e.g. has been used to define the distance between the magnetic tracks in a magnetic disk drive system. 1 fig. 3, the connecting coil 25 represents the basic coil which originates from the connection between the active conductor parts in first winding 2 and second winding 6. However, in addition to the basic coil, the end conductor parts 16- and 17- are also capable of connecting to the end conductor parts 4 in first winding 2. In fig . 3, the connecting rod for the end conductors 16 is represented by the connecting rod 26 (shown dotted). In a similar way, the coupling bolt for the end conductors 17- is represented by the coupling bolt 27 (shown dashed). It appears from fig. 3 that the coupling bolt 26 has a mean value represented by line 52 which is shifted or biased in relation to the mean value, represented by line 54, of the basic coupling bolt 25.

In a similar way, the mean value, represented by line 53, of the coupling wave 27 is also biased in relation to the mean value of the base wave 25.

The coupling of waveform 26 includes an unwanted coupling component which is the constant coupling represented by the offset between lines 52 and 54. Similarly, the offset between lines 53 and 54 represents a constant coupling which is the unwanted coupling component of waveform 27.

According to the invention, the unwanted coupling component of the coupling bolt 26 is equal and opposite to the unwanted coupling component in the coupling bolt 27, so that these coupling bolts, when added algebraically, will neutralize each other. More specifically, the sum of the coupling waves 26 and 27 is in phase with and has zero bias in relation to the basic wave 25.

As long as an unneutralized unwanted coupling component exists, this causes a periodic or fundamental frequency error. The one in fig. The design of the invention shown in 1 neutralizes these unwanted coupling components by the fact that two winding sections are arranged opposite each other in a multi-layer configuration. The constant coupling component in one winding section neutralizes the constant coupling component in the other winding section without affecting

the active leadership parties' basic connection.

In fig. 1, two active leader parties are located over the same room area, e.g. the active leadership parties 18-2 and 19-1. With two active conductors in the same spatial position, a "two-thorn" winding is achieved for the active conductors. However, the end conductors only have "one thorn", as the end conductors in fig. 1 does not occupy the same spatial position. Figs. 4 and 7 schematically show examples of active conductor rows, as they appear at different stages of the manufacturing process for a two-phase, four-layer transformer element as shown schematically in fig. 7. The transformer element in fig. 7 includes it in connection with fig. 1 constant switching (or bias voltage) neutralization features described above. The element in fig. 7 also includes quadrature compensation within each period to ensure that the polyphase sine (BC) and cosine (AD) windings will produce fields that are offset in distance by exactly a quarter space cycle, even when the winding sections making up the sine and cosine windings are not separated exactly in quadrature. Fig. 4 shows a schematic representation of a number of active conductors A, especially A1 to A5. Conductors A1 to A5 are intended to be equally spaced, and the distance between any adjacent pair of active conductor portions denotes half a space cycle. Fig. 4 also includes a number of active conductors B, especially B1 to B5. The active conductors B1 to B5 are separated essentially identically

with the active conductors Al to A5 and is preferably produced from the same photographic negative. The conductors A and B in fig. 4 represents either a photographic negative or isolated metal conductors on a common base of metal, glass or plastic, so that the relative position of the active conductors A is determined in relation to the relative position of the active conductors B. In particular, the active conductors B are displaced an offset "al" from the active conductors A, where "al" is approximately 90 degrees of the space cycle.

Fig. 5 also represents a photographic negative or array of conductors substantially identical to the photographic negative or array of conductors in Fig. 4. Accordingly, the active conductors C, especially Cl to C5, and the active conductors D, especially D1 to D5, are each produced from the respective active conductors Al to A5

and Bl to B5. Consequently, the displacement "a2" between the active conductors D and the active conductors C in fig. 5 essentially identical to the displacement "al" in fig. 4.

In fig. 6 shows a representation of a multi-layer structure, in which the active conductors represented in or photographically developed from fig. 5 are superimposed those from fig. 4 similarly derived active managers. In fig. 6 are some of the active conductors in fig. 4 and 5 removed, and end conductor parts for interconnection of active conductor parts have been added. The additions and omissions are made using well-known printed circuit technology.

In fig. 6, there are four separate winding sections A, B, C and D, whose designations are derived from the designations for the active conductors in fig. 4 and 5. In fig. 6, the displacement a1 between the winding sections A and C is substantially equal to the displacement <x2 between the winding sections B and D. The essential similarity between the displacements a1 and a2 is derived from the identities in FIG. 4 and 5 between the displacements "al" and "a2". When the layers in fig. 4 and 5 are superimposed on each other, the active leader parties B and C come roughly in line. If, as shown in fig. 6, a displacement "b" exists between the "winding sections D and C, al becomes equal to a2, because al = "al" + "b" and a2 = "a2" + "b" and "al" + "b" = " a2" + "b".

The transformer in fig. 7 is produced by rotating the winding sections B and D below the winding sections C and A in fig. 6 so that

the two-layer winding sections in fig. 6 becomes four-layer winding sections in fig. 7. While the winding sections B and D during assembly can be displaced relative to the winding sections C and A, the winding sections C and A and B and D cannot move relative to each other. In fig. 7, the angle (3) between the winding sections B and C can therefore be different from the angle "b" in Fig. 6 between the same sections. The angles a1 and a2 do not vary and are

the same as in fig. 6.

The terminals 32 and 36 of the winding sections A and D, respectively, are electrically interconnected, so that the winding sections A and D form the complete cosine winding with the input terminals 33 and 37.

Similarly, the terminals 35 and 31 are on the winding sections

B and C electrically interconnected, so that they form a complete sine winding with input terminals 30 and 34. Both the sine winding consisting of the winding sections C and B and the cosine winding consisting of the winding sections A and D individually imply that in connection with the winding in fig. 6 previously described inventive features with constant switching neutralization for the same reasons as the winding in fig. 6 does it. Furthermore, the sine and cosine windings in fig. 7 the inventive feature with quadrature compensation which will now be described further in connection with fig. 8.

In fig. 8, each of the vectors A(i), B(i), C(i) and D(i) represents the spatial position of active conductors denoted by capital letters in fig. 7. The lowercase letter "i" represents any of the suffixes from 1 to 5 for the active conductors in fig. 7. For "i" = 2, e.g., the vector diagram in Fig. 8 the spatial positions for the active conductors A2, B2, C2 and D2 in fig. 7. Similarly, fig. 8, for "i" = 3, the spatial position of the active conductors A3, B3, C3 and D3. As stated above in connection with fig. 7, the vector A2 is shifted by an angle al from the vector C2, and in a similar way the vector B2 is shifted by an angle a2 from the vector D2. Furthermore, due to those in connection with fig. 4, 5, 6 and 7 mentioned production steps, the angle al equal to the angle a2. The condition of this equality produces the mutual electrical connection between the vectors B and C, e.g. B2 and C2, a resulting vector BC2 as in fig. 8 is shown generally as BC(i). In a similar way, it produces mutual connection between the vectors A2 and -D2 (-D2 is the electrical inverse of D2), as in fig. 8 is denoted as A(i) and -D(i), a resulting vector AD(2) which is generally denoted as AD(i) in FIG. 8.

With the angle between vectors B and C equal to (3), the resulting vector BC is located (3/2 from both vector B and vector C.

The vectors A and -D are separated by an angle 0 equal to al - [(180° - 2)

+ P], Since al is equal to a2, or more generally equal to a, the angle is given by: 0 = 2a - (3 - 180°. The resulting vector AD is therefore located at a distance 0/2 from both vector A and vector -D. The angle 0/2 is equal to a - p/2 - 90°. As will be seen, the angle

between the resulting vectors AD and BC equal to a - £*/2 - P/2. By inserting the value of 'fj/ 2 given above into the latter expression, it is seen that the angle between the resulting vectors AD and BC becomes exactly 90 degrees. Although fig. 8 is described, for example, with the suffix value "i" equal to 2, the vector diagram in fig. 8 applicable to all suffixes 2 through 4. Since the overall space area over which the active conductors connect with the reference winding for each value of the suffix "i" is less than one space cycle, quadrature compensation is provided individually within each cycle. The use of quadrature compensation within each cycle is important to reduce the sensitivity of the quadrature compensation to anomalies. As the active conductors that contribute to the compensation connect with the same general room areas, compared to widely separated room areas, the quadrature compensation according to the design in fig. 7 relatively immune to the effects of anomalies. In fig. 9 is the resulting vector AD(2), for reasons mentioned in connection with fig. 8, exactly 90 degrees from the resulting vector BC(2). Similarly, vectors A3, B3, C3 and D3 produce resultant vectors AD3 and BC3 which, in accordance with those associated with FIG. 8 principles discussed, are exactly 90 degrees apart. Although the resulting vectors BC2 and BC3 are not exactly aligned, it will be seen that the resultant vectors AD(2,3) and BC(2,3) derived therefrom are also exactly 90 degrees apart. In fig. 10 the winding sections A, B, C and D represent those in fig. ;7 showed. The four winding sections in fig. 10 are, to more clearly highlight their details, separated and not shown stacked on top of each other. In fig. 10, the layers are arranged from the top towards the foot of the page in the order C, A, B and D, which represents their increasing distance from the skaiawinding element in the one in fig. 14 showed a position-measuring transformer. In fig. 10, the winding sections A to D are produced on it in connection with fig. 4 to 7 specified manner. Therefore, when winding sections B and C are interconnected to form the sine winding, and winding sections A and D are interconnected to form the cosine winding, these windings exhibit quadrature compensation. Furthermore, winding sections B and C and winding sections A and D include the constant switching neutralization feature described above in connection with fig. 1, 2 and 3. In fig. 10 bar winding section C several active conductor parts, for which the active conductor 118 is typical. The active conductor parts are interconnected along edges defined by opposite ends of the active conductor parts, of which the end conductor parts 116 and 117 are typical. The center distances of the active leading parties, e.g. the active conductor portions 118 and 119 are equal to the pitch P, and are the same for all winding sections A to D in fig. 10. Although the windings have the same pitch and therefore connect to the reference winding of the position-measuring transformer with the same fundamental space frequency, the widths W of the active conductor sections and the widths S of the spaces between active conductor sections vary from one winding section to another. ;The distance variation of the active conductor parts in relation to the room distances in fig. 10 has two reasons. By "the ratio between the conductor and space distances" is understood here the English expression "conductor-to-space width ratio". First, in accordance with the general teachings of US Patent No. 2,799,835, the ratio of conductor to space distances is used to reduce or neutralize the effects of harmonics in the switching signal. Since the third harmonic element usually makes the largest contribution to error, the ratio of conductor to space distances is usually set to 2:1 for third harmonic neutralization, although other ratios may be chosen for second harmonic neutralization. Secondly, the ratio between conductor and space distances is used to adjust the connection between the winding sections and the reference winding, the winding sections being located at different distances from the reference winding to which they connect. In fig. 14 the different distances for a typical transducer are shown by the dimensions along the left edge. The further a winding section is from the reference winding, the lower the coupling of this section. In view of the purpose of the invention, the coupling is conventionally expressed as a transformation ratio, which is equal to the ratio between the input signal I in the primary winding (e.g. the sine or cosine winding) and the output signal V in the secondary winding (e.g. the reference winding) . In accordance with the present invention, the coupling, and possibly also the resulting transformation ratio, is the lower the greater the ratio between conductor and space distances. According to this principle, the winding sections that are closer to the reference winding have a larger ratio between conductor and space distance, so that their coupling to the reference winding is reduced, which causes the transformation ratio to be reduced and makes this more similar to the transformation ratio for the further away winding sections. In fig. 10, the distance or width of the active conductor portions for the winding sections A, B, C and D is given by WA, WB, WC and WD respectively. In a similar way, the space distance between the active conductor parts for the winding sections A, B, C and D is given by SA, SB, SC and SD respectively. The distances of the winding sections WC and WA for the winding sections C and A are equal, as are the widths WB and WD of the conductor portions for the winding sections B and D. Winding sections C and A are closer to the reference winding than winding sections B and D, see fig. 14. In fig. 12, 13 and 14 show the way in which the winding sections A to D in fig. 10 are placed one above the other to form a multi-layer, multi-phase element 134. In fig. 13, the layers of conductors in windings C, A, B and D are shown progressively cut away from a base layer 136. In fig. 14, the multiphase element 134 is shown below a scale element 128. A typical construction of the multiphase element 134 includes winding section layers B and D applied to opposite sides of an insulating plastic layer 138, and winding section layers C and A applied to opposite sides of an insulating plastic layer 140. The laminate B- 138-D is glued to the base layer 136 by means of another insulating and adhesive layer 137. In a similar way, the laminate C-140-A is glued to the surface of layer B by means of an adhesive and insulating layer 139 between the layers A and B. An electrostatic screen 142 of the type described in US Patent No. 3,090,934 is bonded to layer C by means of an insulating and adhesive layer 141. The scale or reference winding element 128, which is constructed in a manner well known in the art manner, is placed over the multiphase element 134. The element 128 typically includes a continuous reference winding layer 131 which is glued to the base 129 by means of the layer 130. The layer 131 includes a winding of the previously described type which is shown schematically as winding 2 in fig. 1. The thicknesses for a typical multilayer construction are shown in fractions of an inch along the left side of fig. 14. Although these dimensions are not critical, the choice of different thicknesses results in different transformation ratios, which for compensation requires different ratios between conductor and space distance. As discussed above in connection with fig. 10, the ratio between conductor and space distance for the different winding sections is varied in order to compensate for the differences in coupling between the winding sections and the reference winding, which are suitably measured as differences in transformation ratio, and which are principally due to differences in the displacements of the winding sections in the multiphase element 134 from the continuous reference winding in element 128. The variations in the ratio between conductor and space distance shall be further explained in connection with fig. 12. Fig. 12 shows a section of the winding layers along the lines 12-12 in fig. 14, although fig. 12 is a simplified schematic representation which does not show the insulating, adhesive layers or the base and screen layers. Fig. 12 shows pairs of active conductor parts for each of the winding sections C, A, B and D under a reference winding represented by three typical active conductors Ri, R2 and R3. In fig. 14, the conductors Ri, R2 and R3 are included in the reference winding layer 130. The center distance for all the active conductors is equal to the pitch P, and the space cycle is equal to 2P. The distance of each active conductor part in the reference joint is equal to WR, and the distances between adjacent active conductors are equal to SR. In accordance with the teaching contained in US patent no. 2 799 8 35, the ratio between conductor and space distance here is equal to 2:1 for compensation of unwanted third harmonics. This 2:1 ratio is achieved by making WR equal to 2SR. In short, third harmonic compensation is achieved because the distance of WR is equal to one third of 2P. Accordingly, WR is exactly equal to a period of third harmonic space cycle. The average of a sine curve over exactly one period is equal to zero, so that the average coupling for each conductor rod Ri, R2 and R3 will be zero m.h.p. third harmonic. ;For further discussion of the principles of harmonic compensation, reference is made to the latter patent. It is also possible to make the ratio between conductor and space distance for each of the winding sections A, B, C and D equal to 2:1 in order to neutralize unwanted third harmonic coupling. Alternatively, however, said ratio is varied outside of 2:1 in order to more closely equate the transformation ratios for each layer. ;In order to equate the connection to the reference winding and thus the transformation ratio between winding section B and the winding section C to which it is connected, the width of the conductors in winding section C is made larger than the width of the conductor in winding section B. ;In a similar way, the width of the conductors in section A is made larger than the widths for section D. If full equivalence of all winding sections is desired, the widths of the active conductors in section C are made larger than for section A, which in turn is made larger than for section B, which in turn is made larger than for section D In order to simplify the preparation, however, full equivalence is not used in the winding sections in fig. 10. Instead, winding sections C and A are given the same conductor/space distance ratio, and similarly, winding sections B and D are given the same conductor/space distance ratio different from the ratio for C and A. More specifically, for the general thicknesses and displacements as above indicated in connection with fig. 14, found that a width of 0.0393 inch for the conductors and a clearance of 0.0107 inch for the winding sections C and A give good results when the conductor widths are 0.027 3 inch and the clearances are 0.0227 inch for the winding sections B and D. ;For the above stated dimensions in fig. 10 and 14, the coupling for winding section C is of the order of 1.45 times section B's coupling to the reference winding, Similarly, the coupling for section A is of the order of 1.45 times the coupling for section D. To make the coupling for section C equal to the coupling for section ;B and the connection for section A equal to the connection for section D, the transformation ratios of sections C and A are reduced by increasing their conductor/space distance ratios, while the transformation ratios of winding sections B and D are both increased by decreasing their conductor/space distance ratios. ;The necessary change in conductor/space distance ratio can be determined both experimentally and mathematically. For an approximation of the mathematics used to calculate the desired change in conductor/space distance ratio, consider first a winding with a conductor/space distance ratio of 2:1. A reduction of the transformation ratio is achieved by extending the relevant conductor by a size by adding / S./ 2 on each side of the center line, thereby increasing the conductor/space distance ratio. Addition of an additional value on each side of the conductor is an addition at points located approximately 120 degrees from the basic space cycle. The sum of two 120 degree offset vectors is less than the sum of all additional vectors angularly offset by less than 120 degrees. The relevant summation over the full width of a conductor can be done using integral calculus. It is assumed that in the considered area the transformation change for the basic signal is approximately linear. The distance change Z\ which is required to equate the transformation ratios TB and TC which respectively represent the connection of the winding sections B and C to the reference winding, is derived as follows: ;with: [fB/Tc]^ 1/1.45 and ;WC = WB = 0.0333 ;then becomes: KB/KC = 1/1.45 ;To make the transformation ratios equal, i.e. make ;TB = TC, the widths or distances WC and WB are changed by a size ;A, whereby the equation (1 ) can be written: ;Solving equation (2) m.h.p.A , we get: ;0.006 inch ;Where: ;TC = the transformation ratio Vut/Vin, which represents the winding section C coupling to the reference winding. ;TB = the transformation ratio Vut/Vin, which represents the connection of the winding section B to the reference winding. ;WC = the width (0.0333 inch) of the conductor in winding section C, which has a space cycle of 0.1 inch. ;[TB/TC]m = measured value of the ratio TC/TB for the change of magnitude A. ;WB = width (0.0333 inch) of the conductor in winding section B, which has a space cycle of 0.1 inch. ;KC - the proportionality constant for winding section C. ;KB = the proportionality constant for winding section B. ;= width or distance change for WC and WB. Referring to fig. 10, the calculated value for A, 0.006 inch, is added to the nominal 2:1 distance, 0.0333 inch, for winding sections C and A to make WC and WA equal to 0.0393 inch, thereby reducing the clearance distances SC and SA to 0, 0107 inches. Similarly, 0.006 inch is subtracted from the nominal value of 0.0333 to make the distances WB and WD equal to 0.027 3 inches, thereby increasing the space distances SB and SD to 0.0227 inches. Note that the average distance of WC and WB is 0.0333 and of WA and WD is also 0.0333, so that the active conductors in the sine winding (BC) and the cosine winding ;(AD) have an average conductor/space distance ratio of 2: 1... ;The above calculations of A are carried out by principal consideration of the coupling of the winding sections at the fundamental frequency. The effect of the coupling of the third harmonic of the fundamental frequency for delta equal to 0.006 inch is different from the effect for the fundamental frequency. After a change of 0.006, the conductor/space distance ratio in fig. 10 not exactly equal to 2:1, which is why each conductor part does not individually cancel the effects of third-harmonic coupling. Despite this lack of complete third-harmonic compensation, as stated above, the average conductor/space distance ratio for the series-connected winding sections A and D is equal to 2:1. For this reason, any buckling in third-harmonic coupling due to buckling in the conductor/spacing ratios for winding sections C and A will tend to be canceled out by an opposite change in third-harmonic coupling for winding sections B and D, respectively; Since the conductor/space distance ratios are not selected at four different values, the sine winding sections A and C, in order to equate the transformation ratios for all four winding sections in Fig. 10, generally a resulting transformation ratio that is greater than the transformation ratios for the cosine winding sections B and D. Since the quadrature compensation according to the invention generally requires equally large transformation ratios, a further adjustment of the transformation ratios for the winding sections in fig. 10. As shown in fig. 7, this adjustment is made by placing a conventional (not shown) resistance between the sine winding terminals 34 and 30. In this way, the voltage in the sine winding is reduced in relation to the cosine winding in order to compensate for the greater transformation ratio of the sine winding sections. The sine winding has a larger transformation ratio because the sine winding sections C and B, as can be seen from fig. 12 and 14, is closer to an average of the reference winding than the cosine winding-six ions A and D. Using a resistor across one of the windings of a position-measuring transformer is called "line balancing". The present invention ensures, through quadrature compensation, that the sine and cosine windings tend to be exactly 90 degrees apart. According to another feature of the invention, the quadrature can also be compensated by placing a resistor across one or more winding sections. E.g. a resistance between terminals 32 and 33 will reduce winding section A's coupling to the reference winding. Referring to fig. 8, a resistance between terminals 32 and 33 will cause the length of the vector A(i) to decrease, which tends to rotate the resultant vector AD(i) clockwise. If the angle between the resulting vectors AD(i) and BC(i) is greater than 90 degrees, indicating incomplete quadrature, the resistance between terminals 32 and 33 in FIG. 7 tend to correct for the quadrature error. Similarly, for angles less than 90 degrees, a resistance between terminals 37 and 36 causes a counter-clockwise rotation of vector AD(i), which will tend to correct the quadrature error. ;A resistor connected between terminals 30 and 34 or between 33 and 37-in fig. 7 abbreviates the resulting vectors, respectively AD(i) and BC(i) in fig. 8 without affecting the intermediate angle. A resistance between the terminals of any winding section, ie from terminal 32 and either 33 or 37, or from terminal 35 and either 30 or 34, results in a change in the angle between the resultant vectors. Means are therefore arranged to adjust the line balance, i.e. the relative coupling of the sine and cosine windings, and means are also arranged to adjust the quadrature between the sine and cosine windings. Referring to fig. 11, the horizontal conductors 91, 92, 93 and 94 correspond to the horizontal conductors with the same number in fig. 10. Note that the sine winding conductors 91 and 92 are superimposed and electrically connected to conduct in opposite directions to neutralize any unwanted coupling of these conductors. As shown in fig. 11, the terminals 32' and 36' are connected together to connect the two cosine winding sections A and D in FIG. 10. In a similar manner, the terminals 31' and 35' are connected together to connect the two sine winding sections B and C in fig. 10. The terminals in fig. 11 are numbered as the corresponding terminals in fig. 7, but with prime marking ('). In fig. 11 also shows the adjusting resistors 96 and 97. Resistor 96, which is connected between the cosine terminals 32' and 33', effects adjustment of the angle between the sine and cosine windings, i.e. adjustment of the quadrature. Resistor 97, which is connected between cosine terminals 33' and 37', causes adjustment of the current in the cosine winding relative to the sine winding and thus constitutes the "line balancing" resistor. Although resistors 96 and 97 are shown connected between the cosine terminals, resistors can be connected between any terminals in fig. 11 to adjust the relative coupling of any winding section or any combination of winding sections as desired. Although the design discussed in connection with Figures 7 to 14 has windings which all comprise two winding sections which for manufacturing purposes are formed by reversing one section relative to the other, the "reversed" relationship can be achieved if desired without actually reversing the winding sections mechanical. E.g. the end conductor parts can be added by well-known photographic technique along each edge without any mechanical "turning" of the active conductor parts. Figures 15 to 21 show a multi-period rotating transducer. This transducer uses the feature of constant coupling neutralization which is described in connection with figures 1 and 7, the former in connection with fig. 4 to 9 described features with quadrature compensation, and the previously described features in connection with figures 10 to 14 with harmonic cancellation and adjustment of the transformation ratio. ;The monster in fig. 15 schematically shows a series of conductors or alternatively a photographic negative for producing a series of conductors. In particular, the monster in fig. 15 a row A' of radial conductors A10 to A80 which are fixed in space in relation to another row B' of radial conductors B10 to B80. The monster in fig. 15 for rotating devices is analogous to that previously in fig. 4 described monster for linear devices. The radial distance between the conductors in series A<*> is essentially identical to the radial distance between the conductors in series B'. The rows A<1> and B' are preferably produced from the same photographic negative, so that they are identical. Row B' is, however, rotated 180 mechanical degrees plus a quarter of a space cycle for row A'. The space cycle is equal to twice the angular displacement between two adjacent conductors. The rotation of series BT in relation to series A' can be observed by

note the location of conductors AlO and B10.

In fig. 16 shows another monster identical to that in fig. 15,

and comprising a range C of radial conductors ClO to C80 and

a series D' of radial conductors D10 to D80. The monster in fig. 16 is preferably a true copy of the monster in fig. 15 and is produced e.g. by photographic technique. The monster in fig. 16 is rotated 180 degrees in space about an axis perpendicular to the plane of the paper i

compared to the monster in fig. 15. Row D' is preferably a photographic print of row A', and in the same way row C is preferably a print of row B', so that conductor D10 corresponds to conductor AlO and conductor ClO corresponds to conductor B1O.

In fig. 17, a two-layer row is formed by placing the monster in fig. 16 above the monster in fig. 15, so that row Cr lies above and is concentric with row A', and row D' lies above and is concentric with row B'. Fig. 17 for rotating devices is analogous to the previously described fig. 6 for linear devices. In the monster in fig. 17, the angle between conductors B' and D' is identical to the angle between conductors A' and C. This identity is achieved by carrying out the assembly steps described in connection with fig. 15, 16 and 17.

'in fig. 18 schematically only shows conductors with suffix 10 in fig. 17.

The angle "a" between conductors BlO and D10 is equal to the angle "a" between conductors AlO and ClO. When the monster for rows Dr and B' is folded under and stacked concentrically with the monster for rows C and A', which is further described below in connection with fig. 20, any misalignment of conductors A' and D' is indicated as an angle "b" in FIG. 18.

In fig. 18, 20 and 21 show further details of the multi-periodic rotating transducer element. The four-layer monster in fig. 20 is formed by using the conductors from the monster in fig. 17 and fold rows D' and B' under rows C and A' about line 41. In addition to this, for the monster in fig. 20 fbyet end conductor sections, of which sections 62, 63, 64 and 65 are typical, to interconnect adjacent radial conductor sections in an analogous manner to those in fig. 6 provided end conductors. In order to make it easier to see the four separate layers of winding sections, fig. 19 each separate winding section in the series C, A<1>, B<1> and D", where the winding sections are superimposed from top to bottom. Note that the series C, A', B' and D" of the rotating device is the same as for the linear device shown earlier and described in connection with figures 10 and 14. As figures 19 and 20 are schematic, the various insulating and adhesive layers are not shown, but are of course present in the same way as described in connection with Fig. 14.

Fig. 21 shows a vector diagram illustrating the quadrature compensation in the rotary device in fig. 20 in an analogous manner to the preparation in fig. 8 of the quadrature compensation by the linear device in fig. 7. As shown in fig. 21, the resulting vector A'D', representing the resulting spatial position of the cosine winding formed by the cosine winding sections A<1> and Dr, tends to be exactly 90 degrees offset from the resulting vector C'Br representing the spatial position of the sine winding formed of the sine winding sections Cr and B'. In short, the angle 0 between the vectors C and -B1 is equal to (180 - ct2 + (3) - al. Since al = a2 = a, 0 = -2a + |3 + 180, and 0/ 2 -a + p/ 2 + 90. The angle between the resultant vectors A'D' and CD', see Fig. 21, is

a - (3/2 + 0/ 2, which by introducing the value for 0/2 in the previous sentence will be seen to be exactly 90 degrees.

The constant coupling neutralization is incorporated in the design in fig. 20, wherein, also with reference to fig. 19, the sine winding sections-C and B' are connected by a connecting conductor 47 so that their end conductor portions lead in opposite directions. Similarly, the cosine winding sections A' and D' are connected by a strap 46 so that their end conductor portions lead in opposite directions.

More specifically and still with reference to fig. 19 and 20 are

a plurality of end conductor portions (eg, 62) at each end of each winding section located near one of the spaces between the adjacent end conductor portions (64, 68) of the other winding section in the same edge.

In order to incorporate the harmonic neutralization and adjustment of the transformation ratio in the rotary device of FIG. 20, the conductor/space distance ratio is changed in an analogous manner to that described above. In fig. 20 is for an ideal transducer the conductor portions and the space between the wedge-shaped. Because of this shape, both the leader parts and the space between them increase in distance as a function of the distance from the monster's center. For reasons of simplicity in the construction, however, the space between the conductor parts is made rectangular in shape, while the conductor parts are wedge-shaped as shown, e.g. in the above-referenced US patent no. 2 799 8 35. The distance of the rectangular space is made equal to the average distance of a wedge-shaped space. The mean distance of a wedge-shaped space is the distance at the point halfway between the innermost point along a radial conductor and the outermost point. The conductor/space distance ratio is determined for convenience at the point halfway between the inner and outer ends of the radial conductors. In this way, the conductor/space distance ratios for each of the winding sections in fig. 19 is easily adjusted in the same way as described in connection with figures 10 to 14.

The transducer in fig. 20 includes only four space cycles per revolution. In practice, however, transducers with many more active conductors are used to produce a greater number of periods per revolution, e.g. 360 space cycles or more per revolution.

In fig. 22 schematically shows a two-layer multiphase winding 154 in connection with a reference winding 155. The relevant dimensions are typically similar to the dimensions in Fig. 10. The multiphase windings 154 include a sine winding and a cosine winding which are respectively made up of the winding sections B and C and the winding sections A and D. The winding sections B and C lie in the same upper layer and the winding sections D and A in the same lower layer, where these two layers are, for example, on opposite sides of an (not shown) insulator.

The multiphase windings 154 are preferably produced in the manner indicated in connection with Figures 4, 5 and 6. More specifically, it is a monster consisting of row C over row A, as in fig. 6, reversed to get A over C as shown on the right in fig. 22 while the monster consisting of row D over row B is reversed to get B over D as shown on the left in fig. 22. E.g. can the device in fig. 22 is produced by cutting off the monster in fig. 6 along an imaginary line between the monsters C over A and D over B and then turn the latter to get B over D. Then the monsters are glued

C over A and B over D into a common (not shown) base layer to form a laminate in which B and C form a layer over D and A,

which forms another layer denoted by the expression BC/DA in fig. 22. Any misalignment of sample BD with respect to sample CA of the winding sections of FIG. 22 in relation to the reference winding 155 appears as shown as the angle (360 - e.g. between the conductor parts B5 and Cl. In Fig. 22 the displacement a1 between the winding sections A and C is identical to the displacement a2 between the winding sections B and D for reasons which is mentioned in connection with Fig. 6. Therefore, the quadrature compensation mentioned in connection with Fig. 7 is also present in the one in Fig.

22 showed the design of the invention.

In addition to quadrature compensation, each of the sine and cosine windings in fig. 22 separately two continuous winding sections in series to neutralize constant coupling. E.g. the sine winding consists of the winding sections B and C, where the end conductor portions of the sine section B, of which the end conductor portions 56 and 57 are typical, are arranged in a row on opposite edges to lead in the opposite direction to the end conductor portions of the sine section C, for which the end conductor portions 58 and 59, also on opposite sides, are typical. As will be seen from fig. 22 is the connection of the reference winding 155 to e.g. end conductor portions 56 and 57 similar to that shown by curve 26 in fig. 3. In a similar way, the reference winding's 155 connection to e.g. the end conductor parts 58 and 59 in fig. 22 similar to that shown by curve 27 in fig. 3. As described in connection with fig. 3, the constant connection for one set of end conductor parts (56 and 57) is equal and opposite to the connection for the other set (58 and 59). In a similar way, the cosine convolution in fig. 22 also constant coupling neutralization.

In addition to the features mentioned above, the multiphase windings 154 in fig. 22 the previously described features of constant coupling neutralization, harmonic neutralization and coupling (transformation ratio) adjustment. While the four-layer design in fig. 7 and 10 ideally require a different conductor/space spacing ratio for each of the four winding sections to fully equate the couplings (transformation ratios), making the two-layer design of Fig. 22 not this. The requirement for four different ratios led to the four-layer design, because each of the four winding sections is offset to a different degree from the reference winding. In two-layer . the design in fig. 22, where winding sections D and C are in one layer and sections B and A in another layer, ideally only two conductor/space cycle ratios are required to fully equate the transformation ratios for all four winding sections, i.e. to equate the couplings for all four the winding sections of the reference winding.

While the device constructed in accordance with the winding sections in fig. 10 usually requires the connection of a resistor across the sine terminals 30 and 34 of FIG. 7, the device in fig. 22 ideally not a resistor to equate the sine and cosine winding's transformation ratio. However, according to the invention, the connection of the sine-bg cosine windings to the reference winding can be equated and/or the quadrature between the sine-

and the cosine windings can be adjusted by connecting one or more resistors across suitable terminals as described above in connection with the design in fig. 7.

In fig. 22, the terminals of the winding sections A, B, C and D are numbered in accordance with the numbering in fig. 7, but with added prime marking ('). The connections across the terminals in fig. 22

are arranged in series so that they neutralize any unwanted field that might otherwise arise. Referring to winding section A, the vertical conductor 172 above terminal 32' is offset by exactly an integral multiple of the distance period 2P, from the vertical conductor 175 above terminal 33'. Any coupling originating from conductor 172 tends to be neutralized by any coupling of conductor 175, since they each conduct in mutually opposite directions.

If it is desired to make the distance between the vertical parts 172 and 175 different from an integer multiple of 2P, neutralization of the connection of the conductors 172 and 175 can be achieved by means of the connection for the conductors 173 and 174 belonging to winding section D.

In addition to the vertical conductors 172 to 175, horizontal terminal conductors are similarly arranged in series to effect neutralization of unwanted coupling. For example, the conductor 167 connecting the vertical terminal conductor 173 to the terminal connecting means 165 is symmetrical about an imaginary center line between sections BD and AC in relation to the horizontal conductor 168 connecting the vertical conductor 175 to the terminals 164 of the cosine winding. Similarly, the horizontal conductors 169 and 170 which connect the terminal connecting means 166 to the sine winding terminals 163, arranged symmetrically on

each side of said center line.

In fig. 23, curve 162 represents the position error PE for a position-measuring transformer according to the design in fig. 7.

For comparison, curve 161 represents the position error PE for a typical transducer of the prior art. Both curves 161 and 162 were obtained by measuring the position by means of a caliper 72 over the same anomaly represented by the connection 74 between two reference winding sections 76

and 77 after the pusher 72 slides along the axis P from left to right over the connection 74. For the sake of the measurement, one of the sections 76 and 77 had larger air gaps in relation to the pusher 72 in order to emphasize the anomaly at the connection 74. The presence of the anomaly appears as two peaks in curve 161 for the previous embodiment. The error curve 162 for the present invention is comparatively even over all areas.

Figures 24 to 29 schematically show the manufacturing steps used for manufacturing a single-periodic rotary transducer which includes features of the present invention.

In fig. 24 is shown a double spiral monster 176. This monster represents one of four winding sections, the four corresponding to the four sections designated A, B, C and D in the descriptions of previous designs, but not all of which are shown explicitly for the single-periodic bipolar, rotating embodiment. The monster in fig. 24 is located in a rotating position in a direction arbitrarily defined as zero degrees as shown by the pole 179 direction.

In fig. 25 shows a monster identical to that in fig. 24 turned 90 degrees clockwise as shown at pole 181. The monsters in fig. 24 and 25 are representative of the monster in the single-phase section. The two-phase section includes four superimposed and interconnected monsters to form the sine and cosine windings in the manner indicated in connection with Figures 26 to 29.

In fig. 26 a photographic part is shown. Form contains a negative plate 183 on top and a negative plate 184 on the bottom with two double-sided laminates, L1 and L2. Each laminate includes, for example, a first and a second layer of copper, one on each side of an (not shown) insulating layer. Each layer of copper is coated with a photographic film so it can be exposed (printed) by the negatives on the plates 183 and 184. Each of the plates 183 and 184 is provided with a spiral mdnster which in connection with fig. 24 described, and each monster is arranged in series to print in the same direction as shown at pole 192. In particular, plate 183 when properly illuminated will print a spiral monster on the top surface 186 of laminate LI, while plate 184 prints a spiral monster on the bottom surface 190 of laminate L2. During the exposure process in fig. 26, upper surface 189 of laminate L2 and lower surface 187 of laminate LI face each other, and are not exposed. The laminates LI and L2 are then exchanged, laminate L2 being placed on top and laminate LI on the bottom, while at the same time both laminates are rotated 90 degrees with regard to their position in relation to the photographic plates 183 and 184. Furthermore, means (not shown) are used to ensure that the laminates LI and L2 have the same alignment as in fig. 26. For example, two or more removable staples (not shown) may be used through the laminates and negative plates.

In fig. 27 are spiral monsters like the one in fig. 24 for both plates 183 and 184 shown aligned in the same direction again, as shown by arrow 193. Upon exposure, a spiral sample is transferred to upper surface 189 of laminate L2 by means of plate 183, and a spiral sample is transferred to lower surface 187 of laminate Li by means of the plate 184. During the exposure in Fig. 27, the previously unexposed spirals on the inner surfaces 186 and 190 are not disturbed.

In fig. 28, the laminates LI and L2 are represented showing the result of the exposures according to those in fig. 27 described manufacturing steps. Laminate L2 includes a spiral row on upper surface 189 in the direction indicated by arrow 194 and a spiral row on lower surface 190 as indicated by arrow 195 direction, rotated 90 degrees relative to arrow 194. Similarly, laminate LI includes a monster of its upper surface 186 with a direction as shown by arrow 197 and a monster on its lower surface having a direction as shown by arrow 196.

In fig. 29 are the laminates LI and L2 in fig. 28 further displaced by 90 degrees as shown by clockwise rotation of laminate LI in fig. 29 for the assembly of the laminates. As shown

in fig. 29 is the spiral indicated. at arrows 194, 180 degrees out of phase with the spiral indicated by arrows 197, while the spirals indicated by arrows 195 and 196 are in phase. Referring to the above, arrows 194, 195, 196 and 197 represent winding sections of the type indicated above as C, A, D and B respectively. Electrical reversal of winding sections 197 and coupling with winding section 194 produces a sinusoidal winding similar to that in connection with the vector diagram in fig. 21 described. Similarly, the direct coupling of the winding sections represented by arrows 185 and 186 results in a cosine winding, where these sine and cosine winding sections tend to be in exact quadrature.

In addition to the quadrature compensation, the transducer according to Figures 24 to 29 includes constant coupling neutralization as explained in connection with other embodiments of the invention. In fig. 25, winding 180 includes a first winding section in the form of a spiral from terminal 145 to connection point 146. The second winding section extends from terminal 147 to connection point 148. Current in coils 145-146 generally flows in a clockwise direction, while current in coil 147- 148 goes in a counter-clockwise direction. For a transducer reference sample such as that in fig. 25 shown generates the monster in fig. 25 a constant bias of the same magnitude, but of opposite sign for each winding section. For the same reasons as mentioned in connection with fig. 3 represents the constant switching curves 26 and 27 in fig. 3 the coupling for winding sec-

the tions in fig. 25.

In addition to the constant coupling neutralization and the quadrature compensation, the transducer according to Figures 24 to 29 can

use "harmonic neutralization and adjustment of the transformation ratio in the same way as these features are described for other embodiments of the invention. More specifically, the conductor/space distance ratio is adjusted to 2:1 to implement 3rd-harmonic

neutralization. Then the spirals, as in fig. 25, is arranged in four different layers as indicated in fig. 29, the conductor/space distance ratio is also modified from the nominal value of 2:1 to equate the connection.

Although the above invention is described in connection with a spatial phase compensation (quadrature) in a two-phase system, the invention can also be used for three-phase or higher-order systems. For example, a three-phase system includes three windings with 120 electrical degree spatial offsets. For a system with N phases, the transducer comprises N windings, each with a different spatial phase equal to a different integer multiple of 360/N degrees. Each of the N windings is made up of N winding sections selected from a total number of N 2 winding sections, and each section WS(i,j) is indicated by different values of "i" and "j", where "i" and "j" go from 1 to N. The winding sections WS(i,j) are arranged in N monsters which are displaced in relation to each other approx. 360/N degrees, and each monster is identified by a different value of "i" equal to 1, ...... Yet each monster includes N substantially identically arranged winding sections, each identified by a different value of "j" equal to. 1, ...., N. The winding sections WS(i,j) are connected to form the N windings W(l) ..., W(N) with space phases that tend to be exactly 360/N degrees from each other, as follows:

For a three-phase system, the value of N above is equal to 3 and the windings formed are W(1), W(2) and W(3), and these windings tend to be 120 degrees apart.

Although the harmonic neutralization feature is explained in connection with changing the conductor/space distance ratio, all designs of the invention can use other techniques for harmonic neutralization. In particular, one can use the technique of tilting the active conductor parts at an angle different from 90 degrees in relation to the transducer's direction of movement. This technique for harmonic neutralization is, together with other techniques suitable for use in the present invention, described in the above-referenced United States patent no. 2,799,835.

In addition to the techniques described in connection with fig. 22 for neutralizing the unwanted coupling of vertical end connection conductors, an alternative technique consists in having conductors with a width of approx. a period. A one-period wide conductor causes the coupling of half of this conductor to neutralize the coupling of the other half of the conductor, whereby unwanted coupling is avoided.

In connection with all embodiments of the invention, and particularly with reference to the bases 129 and 136 in Figures 13 and 14, the bases on which the winding sections are mounted are insulated, typically of plastic, glass or metal, the latter of which is described in US patent no. .3,202,948.

The coupling between any winding section and any other winding in the present invention can be changed by changing its conductor/spacing ratio. Alteration of the coupling either by series or parallel resistors or by means of conductor/space distance ratios may be used in accordance with the invention, firstly to achieve or adjust a desired quadrature or otherwise, phase relationship between the windings and/or, secondly, to equalize or otherwise adjust one winding's connection to another winding. Although the width or distance of the active conductor parts is discussed in general, the distance of the terminal conductors, for example of those that are perpendicular to the relative movement, can also

is selected to adjust a leader's link to another leader.

For example, superimposed terminal conductors, such as conductors 81 and 82 or 83 and 84 in fig. 10, which is connected to lead i

opposite directions in a multi-layer structure, different distances are given to equalize their connection and thereby neutralize its effect.

Claims (6)

1. Position measuring transducer having relatively movable winding elements where one of said elements includes a continuous reference winding (2), where a second of said elements consists of a first pair (A, B) of polyphase winding sections and a second pair (C, D) of multiphase winding sections, where one winding section in both pairs is in circuit with each other (A+D), and where the other winding section in both pairs is in circuit with each other (B+C), where said first and second pairs of ■winding sections are superimposed with respect to each other in separate layers (C/A/B/D or BC/DA), characterized in that each of the winding sections in each of said pairs comprises a continuous array of evenly spaced conductors. arms, that said first pair (A,B) of multiphase winding sections are placed in approximate space quadrature with respect to each other, that the second pair (C, D) of multiphase winding sections are placed in the same approximate space quadrature with respect to each other, that a first winding section (A) in the first pair of multiphase winding sections is arranged relative to a first (C) winding section in the second pair of multiphase winding sections with an offset a, and that the second winding section for said first pair of multiphase winding sections is arranged relative to the second winding section (D) in said second pair of polyphase winding sections with the same offset oc, where a is the same in the above cases and represents 90° plus an error which is the same in both cases. 2. Transducer as specified in claim 1, characterized in that said first winding sections (C, A) in said multiphase windings are arranged closest to said reference winding (2) and that the distance ratio between conductor and space for said first winding sections (A, C) is greater than the active distance ratio between conductor and space for said second winding sections (B, D) . 3. Transducer as specified in claim 2, characterized in that the average of the distance ratios between conductor and space for said first (C, A) and second (B, D) winding sections for each of said multiphase windings is approximately equal to 2:1, whereby coupling of the third harmonic between said multiphase windings and said reference winding tends to be neutralized. 4. Transducer as stated in claim 1, characterized in that said winding sections (176, 180) have the form of a spiral arrangement. 5. Transducer as stated in claim 1, characterized in that a conductor arm in each of all said winding sections is placed in the same half-cycle space (e.g. A^, B^, C^, D^) 0 6. Transducer as stated in claim 1, characterized in that successive space cycles of said multiphase windings each contain a conductor arm from each of said winding sections.