US2747061A - Means and methods of improving the accuracy and resolution of variable resistors - Google Patents

Description

May 22, 1956 Filed Feb.

OHMS

OHMS

BER 2,747,061

A- P. SOR MEANS AND METHODS OF IMPROVING THE ACCURACY AND RESOLUTION OF VARIABLE RESISTORS 4 Sheets-Sheet 1 I 234 56789|Olll2l3l4l5l6 INTERVALS 0F LENGTH FlG.l

BCX)

IN V EN TOR.

A. PAUL SORBER AGENT l23456789|0lll2l3l4l5$ INTERVALS OF LENGTH FIG. 2

May 22, 1956 Filed Feb. 9, 1953 OHMS VARIATION FROM DESIRED RESISTANCE IN DECIMAL FRACTIONS OF I% A. P. SORBER 2,747,061 MEANS AND METHODS OF IMPROVING THE ACCURACY AND RESOLUTION OF VARIABLE RESISTORS 4 Sheets-Sheet 2 IIOO l23456789l0|l INTERVALS 0F LENGTH FIG.3

I23456789IOIII2|3I4I5I6 INVENTOR- INTERVALS OF LENGTH PAUL SORBER BY F1 5 MW 2,747,061 CURACY May 22, 1956 A. P. SORBER MEANS AND METHODS OF IMPROVING THE AC AND RESOLUTION OF VARIABLE RESISTORS 4 Sheets-Sheet 3 Filed Feb. 9. 1953 FIG] FIG. 6

FIG. 8

INVENTOR. A. PAUL SORBER AGENT May 22, 1956 A. P. SORBER 2,747,061

MEANS AND METHODS OF IMPROVING THE ACCURACY AND RESOLUTION OF VARIABLE RESISTORS Filed Feb. 9, 1953 4 Sheets-Sheet 4 66 FIG. l3 FIG. 10 g 57 l Mg! 5 55 FIG 1: k

62 e4 67 58 60 /El 1 68 SI 63 F 31 INVENTOR. |e.|2 A. PAUL SORBER BY 5& M

AGENT United States Patent MEANS AND METHODS OF IMPROVING THE ACURACY AND RESOLUTIQN 0F VARI- ABLE RESISTORS Adelbert Paul Sol-her, Los Angeles, Calif.; Patti Sorber, administrator of said Adelbert Paul Sorber, deceased Application February 9, 1953, Serial No. 335,947 4 Claims. (Cl. 201-56) The present invention pertains to rheostats and potentiometers, and the preferred species relates more particularly to variable resistors of the helical type.

In a great deal of the electrical apparatus in which resistors are used, it is necessary that they be of a high order of accuracy. Such accuracy has been relatively easy to obtain when the resistances are fixed. This is often accomplished by the accurate positioning of a clamp at the proper position around the resistance wire, such clamp serving as one of the terminals. It has been very difficult, however, to obtain the necessary accuracy with variable resistors. In certain bridge circuits and in electronic computators, as well as in a great deal of other equipment, it is necessary that variable resistors have an accuracy of .05 of 1% and it is sometimes required that the error be reduced to .02 of 1%. resistor of the latter degree of accuracy is one of the primary objects of the present invention.

Another object is to increase the resolution in such devices.

A further object is to obtain quicker heat dissipation.

Still another object is noise reduction.

Yet another object is to provide variable resistors having greater uniformity in total resistance.

An additional object is to provide a structure in which desired departures from linearity can be achieved with greater accuracy than in the past. It is sometimes desirable to produce a variable resistor in which the taper might be elliptical, hyperbolic, exponential, or of some other form, and the present invention is intended to facilitate the manufacture of resistors of such type.

Still another object is the provision of a method to facilitate the manufacture of variable resistances having a linear taper.

Additional objects cation proceeds.

In the figures:

Figure 1 is a graph showing the measured values of resistance obtained at progressive intervals along a resistance element as compared to absolute linearity in the ohms and length relationship.

Fig. 2 shows the measured taper of the resistance element whose taper was compared with ideal values in Fig. 1, and the other curve in Fig. 2 shows the taper of another resistance element in which the variations from linearity are generally opposite from those shown in the first mentioned curve.

Fig. 3 shows the average resistance for given intervals of length for the two resistance elements whose values are plotted in Fig. 2.

Fig. 4 is a different form of graph in which the zero center line represents the ideal resistance for various intervals of length. These values may be those required for a variable resistor having a linear taper, or they may represent the calculated ideal values for a taper of hyperbolic, exponential or other form. The upper curve shows the measured deviations from the desired values in one resistance element and the lower curve shows such variawill become apparent as the specifi- To provide a variable tions for another resistance element. is the average of the two.

Fig. 5 is a schematic diagram showing how two resistors may be placed in parallel and used with a double wiperp Fig. 6 is a longitudinal section of a circular rheostat or potentiometer of a conventional form in which two resistors have been placed in parallel-both electrically and physically.

Fig. 7 is a right end view of the device of Fig. 6.

Fig. 8 is a longitudinal section of a helically wound potentiometer employing two helices.

Fig. 9 is a cut-away view of a helical potentiometer in which three helices are placed side by side.

Fig. 10 is a section taken on line 10-10 of Fig. 8.

Fig. 11 is a section taken on line 11-11 of Fig. 8.

Fig. 12 is a top, or plan, view of the wiper and support shown in section in Fig. 11, but with the wiper somewhat extended in length as is done in common practice.

Fig. 13 is a detailed perspective view of the spring contact employed in the device of Figs. 8 and 11.

When variable resistors of a high degree of accuracy are required, it is frequently possible to find individual units that come within the required tolerances even though the variable resistors in the group from which the selection is made may all have contained resistance elements that were not measured prior to assembly; but it will readily be understood that a much higher degree of accuracy is obtainable if the resistance of the elements is measured at various places throughout their lengths before the devices are put together. The present invention makes it possible to produce variable resistors Without such preliminary measurements that nevertheless have a much higher degree of accuracy than has heretofore been obtainable without preliminary tests.

If one resistance element has a resistance variation approaching the maximum permitted deviation within a given portion of its length, it is not likely that another similar resistor selected at random would exhibit an equally great deviation in the corresponding portion of its range. Therefore, if two resistance elements, both wound by the same method from Wire having the same tolerances, are placed side by side and swept by a common wiper, or by two wipers interconnected, the magnitude of any variation from desired values is likely to be reduced. For example, if the variation from the desired value in a given portion of a variable resistors range is of the order of 1%, it is not likely that another resistance element, wound by the same method, would have as great a variation within the same portion of its range. As a The center curve consequence, it two such elements are placed in parallel,

the resistor having the smaller variation serves to bring down the average variation of the two. Moreover, the: chances are one out of two that Where the variation in one resistor is of a positive value the variation in theother resistor will be of a negative value, and the variations will consequently tend to cancel out. For these reasons, it is possible to use two resistance elements in parallel with wipers also connected in parallel and thereby reduce the average variation from the ideal value by one-half.

If the resistance elements are measured before their assembly at various intervals throughout their length, it is possible to select individual resistance elements that approach much closer to the required taper than would be possible without such preliminary measurements. But when the structure of the device is such that two or more resistance elements may be placed in parallel, it is possible greatly to reduce the error by selecting resistors Whose variations throughout their length are of opposite value.

In Fig. 1, curve A shows the taper of a resistor that was measured at 16 diflerent intervals. For a resistor that is to have a high order of accuracy, it is desirable to make far more measurements than this graph indicates, and it is common practice to measure a resistance element of 70 to 100 inches in length at intervals spaced by only one inch: Fig.1 nevertheless serves'to illustrate the principle involved. It will be noted that curve A varies from curve B, which indicates theoretical perfection, by not more than 50 ohms at any given point throughout its entire length and that the total resistance is approximately 1600 c-hms. This is a variation of a very high order and would only be permissible in variable resistors where accuracy is not required. But a resistor having far greater accuracy can be assembled merely by using two selected resistors in parallel, care being taken to see that large variations in one resistor are offset by corresponding variations in the opposite direction in the other resistor.

Curve C in Fig. 2 shows the measured taper of a selected resistor whose resistance variations are generally opposite in value from those of the resistor whose taper is shown by curve A.

In Fig. 3 a single curve AC is shown, this curve representing an average of the two shown in Fig. 2. it will be noted that the variations have been very greatly reduced and that the maximum variation, which occurs at line 7, is a variation of only about 2%) ohms, as compared to a maximum variation of 56 ohms in each of the two resistors when considered alone. Two such resistors connected in parallel and swept by electrically connected wipers, all as indicated in Fig. 5, thus produce a variable resistor of much greater accuracy than can be made with single resistors having the same individual tolerances. Each of the resistors to be thus mated must of course have twice the resistance that is required of the two when connected in parallel.

Fig. 6 shows in longitudinal section a single turn, or circular, variable resistor in which two resistance elements 1 and 2 are employed. These are cemented or otherwise suitably retained in grooves 3 and 4 that are cut or formed in the non-conductive ring-shaped member 5. Member 5 is attached to the non-conductive cup-shaped element 6 by means of screws 7. An operating or control shaft 8 passes through opening 9 in the cup-shaped supporting member 6. Collars 1i and 11 constrain shaft 3 against longitudinal movement. A non-conducting sleeve 12, having a flange 13 at one end, is mounted on shaft 2 with its flange in contact with collar 11. The conducting member 14 has a large opening in one end and a smaller opening in the other, and the large aperture is slipped over the small end of the non-conducting sleeve 12, and member 14 is positioned against flange 13 of sleeve 12 in the manner indicated in Fig. 6. A conducting screw 15, extending through the smaller aperture in member 14, serves to electrically connect member 14 with terminal 16, and the nut 17 holds the lug and member 14 in assembled position. The U-shaped member 18 is formed of resilient metal and constitutes a double wiper. This U-shaped wiper has holes in both arms so that it may he slipped over the smaller portion of member 12' and positioned in contact with the conducting element 14. A nonconducting washer or ring 19, having an inside diameter greater than the outside diameter of the smaller portion of the non-conducting sleeve 12, is placed over the small end of this sleeve.

The threaded end 20 of shaft 18 extends from the small end of sleeve 12, and nut 21 is tightened on this threaded end so that it presses against ring 19, which, in turn, presses on the resilient wiper 18 and locks the various parts surrounding shaft 8 in assembled position. Termi rials 22 and 23, Fig. 7, are held firmly against the opposite ends of resistance element 2 by the mounting screws 24 and 25, and it will be understood by those skilled in the art that similar clamping members, which do not show in the figures, are associated with resistance element 1 and that they are electrically connected to the terminals 22 and 23. Such connection may be provided by the mounting screws 24 and 25.

When the control shaft 8 is rotated, the resilient contact or wiper 18 sweeps over the wound resistance elements in the manner well known in the art.

Aside from the illustrated use of two resistance elements in parallel, swept by a common wiper, or by interconnected wipers, no patentable novelty is claimed in the structure shown in Figs. 6 and 7, and further structural details will therefore not be described. It will be understood, however, that any other structure employing a plurality of resistance elements connected in the manner described and shown, will nevertheless embody the invention that is the subject of this specification.

In helical potentiometers it is possible to use resistance elements of great length, as the helical formation makes it possible to keep the over-all dimensions within reasonable limits. Fig. 8 is a longitudinal section through a potentiometer embodying my invention and containing two resistance elements, each formed into a helix and threaded together so that corresponding portions of the two helices are in juxtaposition. The left ends of these two helices 26 and 27 are both in contact with the clamping member 28, which is mounted on the left end member 30 of the potentiometer by means of the screw 29. This screw also serves to mount the terminal 31 and to connect it electrically with the bracket 28. A similar bracket 32 engages the right ends of helices Z6 and 27, and bracket 32 is mounted on the right end member 33 of the potentiometer by means of screw 34, which also serves to mount the terminal 35 and to connect it electrically with bracket 32 and the right ends of the two helices.

Although corresponding convolutions of the two helices may appear in the drawings to be in contact, i prefer to employ intervening insulation. Air may be used, or the convolutions may be spaced by a web of insulating material, or the resistance wire itself may be enamel covered and the enamel later removed where contact is to be made with the wipers.

The wipers, or sliding contacts, are carried by an insulating support 36, which is slidably mounted by means of a slot 36' in the bottom thereof upon a plate 37 that is centrally disposed within the helices. This plate is attached at opposite ends to shafts 38 and 39, each of which has a central slot to receive one end of the plate, as shown in Figs. 8 and I0; and the plate is soldered, welded, or otherwise rigidly attached to these short shafts. A spacing washer 40 surrounds shaft 38 and is disposed against the left end of plate 37. The free end of shaft 38 is journaled in the end piece 39. In assembly, an insulating. disc 41 with a centrally disposed boss and having an aperture extending both through the boss and through the main body of the disc, is slipped over shaft 39, as indicated in Fig. 8. A centrally apertured conducting wafer 42 is placed over the boss portion of member 41 so that it lies against the left vertical edge of this disc. The portion of shaft 39 that extends from the boss of disc 41 is threaded, as indicated in the figure. A non-conducting ring 48, having an inside diameter slightly larger than the outside diameter of the boss on disc 41, is slipped over this boss; and the parts are of such relative size that this ring extends out from the boss and over the threaded portion of shaft 39. The nut 43 is threaded onto this portion of the shaft and tightened against ring 48 in order to hold the conducting wafer or slip ring 42 firmly in position. The reduced outer end of shaft 39 is housed in bearing 44, which has a flange that fits against the bottom of hole 45 in the non-conducting end plate 33. An internally threaded member 46, whose internal threads are in screw threaded engagement with the external threads on hearing 44, holds bearing 44 in assembled'positio'n.

A flat headed screw 47 extends through slip ring 42 from its face, and it then passes through non-conducting wafer 41 and through the terminal lug 49. A nut 50 holds this lug firmly in position and draws the head of screw 47 tightly into the COlfn'ltEiI'SllDk recess in which it fits in slip ring 42. Conductor 51 electrically connects lug 49 with screw 52, which extends through plate 37 but is insulated therefrom in the well known manner illustrated in Fig. 11. A conducting sleeve 53 surrounds the central portion of the screw 52 where it passes through plate 37, and insulating washers 54 and 55 insulate the lug 56 and the nut 57 from the plate.

A resilient wiper or contact member 58, shown in Figs. 11, 12 and i3, is mounted by means of screws 60 and 60' within the recess 59, Fig. 8, in the top portion of support 36. This contact member is bent back over itself in U-shaped fashion, as clearly shown in Figs. 11 and 13, and the part that extends beyond member 36 is divided to form two independently-acting spring portions 61 and 62, which respectively carry contact members 63 and 64 near their free ends. These contacts are formed of silver or other suitable contact material to keep arcing at a minimum. The spring portions 61 and 62 not only serve to hold the contacts 63 and 64 yieldingly against the helices, but their resilient action also keeps the support 35 in position on plate 37. A lug 65 is mounted under the head of screw 66, and a conductor 66 electrically connects this lug to lug 56. It will thus be clear that the two contact elements 63 and 64 are in electrical communication at all times with slip ring 42 by means of the electrically interconnected parts hereinbefore described.

The contact supporting member 36 has two upwardly extending tines 67 and 68, both of which can be clearly seen in Figs. 8 and 12, but only tine 67 may be seen in Fig. 11, as the line along which this section is taken extends between the tines, as shown in Fig. 8. Tines 67 and 68 ride in the spaces between the paired convolutions of the helices, as shown in Fig. 8. It will be noted that these tines or guides do not extend into a space between adjoining turns or convolutions of the same helix, but that each tine extends between the right side of one of the convolutions of helix 27 and the left side of the adjoining convolution of helix 26. When the control shaft 39 is rotated, tines 67 and 6S, riding between the convolutions of different helices, guide the sliding contacts 63 and 64 along the resistance elements with which they make sliding contact at all times.

it is not essential that two upwardly extending members or tines such as those designated by the numerals 67 and 68 be provided on the support 36, as it will be clear that one such tine will alone serve to guide the contacts 63 and 64 respectively along helices 27 and 26 while the supporting plate 37 is rotated.

A wiper or brush 6%, Fig. 8, is supported by screw 76 and nut '71 on the inner surface of the right end 33 of the potentiometer assembly, and the free end of this flexible brush is urged into sliding contact with the face of the slip ring or wafer 42. A terminal lug 72 is mounted under the head of screw 70, and this lug is therefore in electrical communication at all times with the brush 69 as well as with the other electrically interconnected parts that electrically interconnect the brush 69 with the contacts 63 and 64 on the flexible end of wiper 53.

The helices are cemented or otherwise suitably attached to the inner wall 73 of the cylindrical case 74 of the potentiometer assembly.

it will be noted in Fig. 8 that the shafts 38 and 39 are not symmetrically disposed with respect to the two sides of member 37. This is an intentional departure from symmetry so that the weight of the support 36 and of the sliding contact will be ofi'set by the greater amount of material lying on the opposite side of the rotational axis of plate 37. In potentiometers of this type the rotors are often spun very rapidly until the desired angular position is reached on a given convolution of the helix, such locations being pre-set on the dials of the apparatus in which such variable resistors are used. In other applications, the rotors spin rapidly until the exact amount of wire is in circuit to produce a desired value of resistance. In either case, it is important that the rotating parts be balanced to prevent excessive wear and that the parts be kept light in order that momentum may not cause the wipers to move past the required locations on their respectively associated helices.

Helical potentiometers have been constructed with from two to very large number of convolutions having a total length of well over inches. When these wound resistance elements that are formed into helices are first electrically measured in order to determine the variations from the desired values of resistance at different points through their length, it is possible to pair them up so that wide variations of resistance in a given portion of one helix are minimized or neutralized by lesser or opposite variations in the corresponding portions of the selected companion resistance element. As previously mentioned, it is a common practice to measure the resistance of these elements at numerous points throughout their length, more than one hundred readings having been taken for individual elements, and it is also common practice to use equipment that plots a continuous curve of the resistance variations over the entire length of the elements. The variations indicated in Fig. 4, as hereinbefore explained, show actual measured variations from desired values in two diflferent resistors whose variations are indicated respectively by the upper and lower curves. When such resistance elements are properly selected they can be paired so that the variations effectively neutralize each other, as shown by the central curve in Fig. 4-. Helices for the potentiometer assembly shown in Fig. 8 should be formed from resistance elements that have been thus paired.

it has previously been mentioned that more than two resistance elements may be selected so that variations are still further reduced, the greater number tending to hold the resistance variations to an average value that approximates the ideal. in most applications, the use of a plurality of resistors that have been paired in the manner hereinbefore described will result in a much closer approach to linearity than has heretofore been possible.

My invention also makes it possible to produce rheostats and potentiometers having tapers of non-linear form that more closely approach the required values than have the non-linear variable resistors of the prior art. Variations in the taper for any given resistance element can be produced by varying the spacing of the turns of the resistance wire on the insulated wire form on which such resistance wire is wound; and if such non-linear resistance elements are paired so that variations from the desired non-linear taper in one element are at least partially minimized or neutralized by variations in the other, any desired taper can be more nearly approximated than has heretofore been possible.

As stated near the beginning of this specification, the objects of the invention hereinbefore disclosed include an improvement in the resolution of variable resistors. Resolution, of course, increases with the number of increments of change. If. the wiper covers an average of five turns of the fine resistance wire that is wound on the insulated wire core later formed into a helix, an increase to ten equally spaced turns would double the number of increments of change and thus effect a 100% improvement in the resolution. When two helices are used, no increase in resolution would be obtained if the advancing edge of each wiper engaged a new turn of resistance wire at precisely the same instant and also if the opposite or trailing edges of the two wipers simultaneously broke contact with the turns being left behind, but only rarely would the turns of resistance wire on the two helices be in such perfect juxtaposition that this situation would prevail. If the turns of resistance wire on one heliX were always in uniform staggered relationship with respect to the turns on the adjoining helix, the two wipers would alternately engage approaching turns and the increments of change would thus be effectively doubled; but this ideal situation would also seldom prevail. The real situation will usually be somewhere between the two extreme conditions just described-that is, between a rare condition in which the increments of change with respect to the two wipers are simultaneously eitective and consequently produce no improvement in resolution and the equally rare opposite condition wherein the increments of change with respect to the two wipers are alternately effective and therefore produce a 100% increase in resolution. The average net improvement in resolution will therefore be approximately 50%.

The same factors that produce improvement in resolution, also contribute to noise reduction when variable resistors embodying my invention are substituted for other wire wound resistors in any kind of audio equipent. The fact that in wire wound variable resistors the change in resistance is effected in increments rather than by step-less variation, is itself productive of noise; and noise thus produced is of course reduced by my invention, because the increase in the number of increments of change produces a corresponding decrease in the magnitude of change due to each increment, and the resulting current variations that cause noise are consequently of lesser magni ude. The increase in the number of increments thus provides a smoother control of resistance with a resultant reduction in noise.

The use of two wipers is also helpful in this regard. The extra wiper decreases the possibility of situations arising in which a loose contact between a wiper and t c resistance clement rapidly opens and closes the circuit, thus producing the familiar crackling and static-like sounds that accompany such loose contacts. When two wipers are used, there is little likelihood. that both will be disengaged at the same time; and if one contact is tight while the other is loose, the makes and breaks caused by the loose contact merely result in resistance changes instead of a circuit that is rapidly opening and closing.

It was also mentioned in the enumeration of the obiects that the present invention is intended to provide an increase in heat dissipation. The combined periphery of two resistance elements is of course double that of a single element of the same diameter and there is a consequent increase of 100% in the heat radiating surface when two slightly spaced resistance elements are used. In order to produce the same total resistance with two elements in parallel, that would be obtained when using only one, the required doubled resistance in each of the two individual elements does not always result in the use of elements each having the same periphery that a single element of suitable resistance would have, and it therefore cannot be assumed that the present invention always doubles the heat dissipation; but the proper use of two helices greatly increases the heat radiating surface and consequently appreciably improves the dissipation.

The variable resistors shown and described in this specification are merely illustrative embodiments of my invention, and many other suitable structures may be made of widely varying form, that will carry out the teachings herein set forth. Parts may be added or omitted, and any of the elements of the combinations recited in the appended claims may be replaced by other elements that perform equivalent functions without dcparting from the broad spirit of my invention.

I claim:

1. In a variable resistor, a plurality of resistance elements disposed side by side throughout their lengths, said elements selected and arranged so that departures from the desired resistance in any unit of length of one element are at least partially compensated for by opposite variations in resistance in the corresponding unit of length of at least one other resistance element, a plurality of wipers each individual to one of said elements, and means common to all of said Wipers for moving them simultaneously along the respectively associated elements.

2. In a variable resistor, a plurality of resistance elements disposed side by side throughout their lengths, said elements selected and arranged so that departures from the desired resistance in any unit of length of one element are at least partially compensated for by opposite variationsin resistance in the corresponding unit of length of at least one other resistance element, a plurality of wipers each individual to one of said elements, a first electrical terminal connected to one end of each of said elements, a second electrical terminal connected to each of said wipers, and means common to all of said wipers for moving them simultaneously along the respectively associated elements.

3. In a variable resistor, a plurality of resistance elements of substantially equal total resistance disposed side by side throughout their lengths and formed into a multiple helix, said elements selected and arranged so that departures from the desired resistance in any unit of length of one element are at least partially compensated for by opposite variations in resistance in the corresponding and adjacent unit of length of at least one other resistance element, a plurality of wipers each individual to one of said elements, a first electrical terminal connected to one end of each of said elements, a second electrical terminal connected to each of said wipers, and means common to all of said Wipers for moving them simultaneously along the respectively associated elements.

4. The method of producing a variable resistor which includes measuring the resistance of a plurality of resistance elements between corresponding points throughout their lengths, selecting from the resistance elements thus measured a pair having corresponding sections of length that vary oppositely from the desired resistance, and so mounting said elements on a common support that such corresponding sections of length are disposed adjacently.

References Cited in the file of this patent UNITED STATES PATENTS 733,610 Yates July 14, 1903 1,449,725 Becker Mar. 27, 1923 1,609,846 Troy Dec. 7, 1926 1,883,098 Templeton Oct. 18, 1932 1,917,675 Weichelt July 11, 1933 2,442,469 Palya June 1, 1948

Images