AT MICROFIGHE REFERENCE LIBRARY A project of Volunteers in Asia by: R.H. Warring Pubiished by: Model and Allied Publications Argus Books Limited P.0. Box 35, Wolsey House Wolsey Rd., Hemel Hempstead Hettfordshire HP2 485 England Paper copies ace § 2.25. Available from: META Publications £.0. Box 128 Harblemount, WA 98267 USA Reproduced by permission of Model and Allied Publications. Reproduction of this microfiche document in any form is subject to the same restrictions as those of the original document,fe rad i NaN rons > TECHNICAL PUBLICATIONSpring Design and Calculation R. H. WARRING Model & Allied Publications Limited 13-35 Bridge Sweet, Hemet Hempstead, Herts, EnglandMade! & allied Publications Lia Book Division Staon Road, Kings Langley Henfordshire, England Furst Published 1973 RH, Warting 1978 ISBN © 85242 327 6 Printed and made in England by Page Bros (Norwich? Lid, Norwich CONTENTS 1 Spring Materals 2 Simple Flat Springs Hetical Springs Tapered Helical Springs 3 4 § Torsion Springs 6 Clock Springs 7 Constant Force Springs & Muliipte Leaf Springs Appendix 4 Spring Terminology Appendix B Wire Sizes and Values of d* and d+ Table | Spring Materials and their Mechanical Properties Table {I WahI's Correction Factor K for Round Wire Heleai Coil Springs Table [IF Correction Factors for Rectangular Wire Helica! Coil Springs Table Iv Stress Correction Factors for Torsion Springs Table ¥ Design Vatues for ‘Tensator’ Springs. 39 39 39 39SPRING MATERIALS ~ 4 The spinginess of mecals is slated in a general way to their hardness. Lead, for exampic, is 2 sof: etal, with virtually no ‘spring’ properties. The same with aluminium. cxtrerne hardness, 2n the other nand, again results in lack of ‘s-1ing’ properties hecalise the materiat is brittle cather than ‘alastu:. The cance of surtable spring materia's are thus those which combine seitable iardness with ‘elasticity’ I1:s also importa, if spring ne-tormanre is to be cansistent, that the material retans its orpinal properties. Many metals are subject to ‘work-hardeniny’ of 4 change of hardness when sticssed — and all working springs ate subject "a nycles xf strass. B-ass, for example. is a etal which is selatively sofi, but repssted siressing or ‘working’ causes mts hardness *v incr.ase, w'. tir, metal becoming mare springy as a consequence Thus veniist soft brass is quite useless as a spring matetial, fully hardened bars messesses reasonably good spring prore.ues. The hardiness of many metals can 2180 be improved Ly heat treatment and 25 3 resu't their spring properties enhanced. This ig quite common practice aw The preparation c/ basic spring ‘stock’. Heat, however, can also prudure the opposi's result. Thus a hard, sptingy metal can often be oermanently softered by heating and slow cooling (or annealing). Gri the other hand, heating and tapid ccaling a spring material can increase its hardness ta ihe point of brittleness. Without considerable experiance in the .echniques af heat treatrient therefore, spring materials should always be used ‘as is’. With suitable knowledge. however. processed springs may often be heat treated to advantage - €.9. to remove ir:-ernal stresses remaining in the niteriat after cokd working to shape cr form. The temperature and methed of heat treatv-ent employed is dependent on the com- position of the spring matesial and the method of spring application. Another form of treatment which can produce embrittlement in a spring is electroniating. This applies particularly in the case of carbon steel springs, where plating may sometimes by thought desirable to provide resistance to corrosion. If such springs are plated, regardless of the method used they require to be baked immediately after plating 10 Grive out hydrogen ahksorbed by the materia’ during plating, Any hydrogen temaining in the pores of the spring material will cause embrittement. Sinaiar comrent apolies to plated steel wire, used as a spring material. The renge of true spring materials is fairly limited. Ordinary carbon steel rendered in “spring temper’ form is the mast common choice for general purpose springs of all types. In the case of wire, the necessary temper may be produced by the method of fabrication—e.g. cold aSPRING MATERIALS drawing. The spring temper may. however, be further improved by heat treatment or oil tampering. Such spring materials are suitable for use under ordinary temperatures, in normal stress ranges — i.e. without the limit of proportionality of the material (see later}. For use under higher stresses, o: higher temperatures, special alloy steels may be needed Where corrosion may 5¢ a problem, the choice of stainiess steel or non-ferrous spring materials may be necessary — the former where high stresses have to be carried by the spring, afd the latter for lower cost, easier working, where stresses are not sO high. Berylium copper is an attractive choice where high resistance to stess and corrosion are necessary, and good electrical conductivity is also required. If electrical conductivity is the main requirement, phosphor-bronze provides a cheaper alternative: and brass even lower cost (although brass is a ‘marginal’ spring material, even at full ‘spring temper’). A nickel alloy {e.g. monel) may be specified where high temperatures have to be accommodated. Mechanical Considerations However good the spring material, there are Hmits over which it can be expected ta work consistently and show a long ‘spring’ life. The critical factor involved is the actual stress born by the material when Fig. t limit af prapeetignelity General sofe working range 100% limit oF probartionality)SPRING MATERIALS the spring is working. Ua te a certain point, with increasing stress the corresponding strain in the materal fotows @ linear retationship — Fig 1, Beyond this fimit of proportionality this linear relationship no lenger =pplies and subjecting the material to these higher stress values may permanently change the mechanical sroperties of the material The limit of propertionaliy thus represents the upper stress limit for the material In practice, 2 lewer lint is normaliy employed — 80 per ent of the limit of proportionality — to allow a safety factor in spring design. Working within this limit will then énsure a consistent performance from 2 spring materia This. however. only presents part of the picture. The strength of any material is different for different ways in which it is stressed. Maximum strength is usually available when stressed in pure compression, with an almost similar value when stressed in pure tension. If subject to twisting of Jorsian, the material strength available is considerably reduced Basically, in fact. the lite of 4 spring depends on four main tsctors: (i) The manner in which the spring material is stressed Gi) The maximum warking stress (hii) The range of sess over which the spring material is worked {iv} the number of cycles of stress or the effects of fatigue on maternal properties. Items (i) and (ii) are directly related. Once the manner in which the material is sessed is established, a safe maximum working stress can be established for a particular material — see Table |. The stress range is more difficult to establish. In general, the higher the range of stress over which the spring is worked. the lower should be the maxirnum permissible stiess to ensure long spring life. However, this will vary with both differences in material properties and heat treaiment. and with frequency of working. For simplicity of design best to adept ‘safe’ figures which err on the side of underestimating material performance. such as given in Table I, Whilst maternal strength and strese determine the Iaad which can be carried by a spring of given geometry, and the life of the spring, detlaction characteristics are determined by the moduli of the matazial. Again this depends on the manner in which the spring material is deflected or stretched. If the spring material is under tension, then it is the modulus of elasticity of Young's modulus which is the parameter involved. For a spring material subject to torsion it is the moduius of rigidity which is involved in caiculating deflection. Values of modulus of elasticity (E) and modulus of rigidity (G} have. therefore, to be known for the spying materials used befere the full sSPRING MATERIALS performance of a spring can be evaluated. These are also given in Table |. The modulus of elasticity largely governs the matevial perform- ance of flat springs and torsion springs. The modulus of rigidity governs the material performance in helical springs, The actual stress praduced in @ spring, on the other hand, is dependent only on the load cabried by the spring and the spring geometry, All these individual parameters appear in the spring design formulas in subsequent chapters. Formulas and Units Spring design proportions aré not something thal can be ‘guesstimated’ with any degree of accuracy — and tial-and-error design can produce a succession of failures. Thus this book on spring design is full of formulas, as the only accurate method of predicting soring performance. However, all are essentially practical working formufas, and all are quite strdightforward to use. Each calculation is nathing more elaborate than an arithmetical calculation ~ aided by a slide rule or log tables No units ae given with the formulas, since these follow quite logically depending on whether you are working ta English or metric standards. Most quantities are linear dimensions, and itis only necessary 10 remember that stess values, ete. should be tendered in the same units, Thus tor working with alt dimensions in inches, stresses, ete., must be in pounds per square inch. Answers will then work out logically in the right units. For example, the deflection per coit of a helical compression spring is given by Gat where P is the load Dis the mean coil diameter dis the wire diameter G is the modulus of rigi deflection = ty of the spring material In English units, P would be in pounds. Dimensions D and d would be in inches. To be consistent. G must then be in pounds per square inch. The deflection, calculated from the formula. is another linear dimension and so would be given direetly in inches. Using metric units the point to watch is that the modutys or stress values used (or Calculated) are in the Same units as the linear dimen- iOS. The latter, for example, will usually be in millimetres. Moduli and stress figutes may. however, be quoted in kilagrams per square centi- metre and would need adjusting for consistency when used with millimetre linear units.SIMPLE FLAT SPRINGS 2 The basic form of a simple flat spring is shown in Fig. 2. The following are the twe design formulas concerned: 6PL 4PL> — flection = 7 bE Deflection a in consistent units — ie, Swess is given in Ib/sq.in when P is in pounds and Lb and 1are in inches. The modulus of elasticity (E} is in Ib/sq.in. Stress (tension) Fig. 2 As a general guide it follows that: Stress in the spring material increases in direct proportion to spring length, and in inverse propertion to width and ¢thickness)?, Thus, tor example, increasing the thickness of the spring will decrease the stress more effectively than increasing the width Deffection increases with the cube of the length (thus small changes in length will have a marked effect on deflection); but can be decreased by increasing the witith or thickness (the latter being much more effective in stiffening the spring). Design Procedure The design of a simple flat spring commonly calls for a certain deflec- tion not to be exceeded under a given load. The length (L) of the spring may also be predetermined, or can be given a suitable value. The formula for Deflection can then be reerranged as a solution for bi, viz ae A = —______ bt ~ Ey deflection All values on the right hand side of this equation are known (the value of E following from the spring material selected), and can thus be calculated as a single quantity — say X ?SIMPLE FLAT SPRINGS. We then have be =x From this point, either ‘guesstimate’ a value of b and from this calculate the cemesponding value of t to satisfy the equation; or ‘guesstimate’ t and from this calculate b. The latter is the usual method since thickness is gaverned by the standard sizes of materials availabie, and thus there is a choice of specific values of t (e.g. 20 swg, 18 swg, etc.), Note: See Appendix B for tabular values of 17. Any solutions derived by the above method will give spring propor- tions satisfying the deflection under load requirements. |i is now necessary 10 enter these values in the Stress formula. together with loag (P) and calculate the stress resulting. Providing this is lower than the maximum permissible stress for the material used, then the spring geometty is satistactory. If the calculated stress is Aigher than the maximum permissible stress. then the spring geometry must be recalculated from the deflection formula. using different values. This is simplest i the spring length is left unaltered. It is then only necessary to return to the formula br? = x and use a greater thickness to calculate a new value for b. Check if this reduced the stress to below the maximum permissible value. |f not. try again. Square Wire Springs Use the same farrtuias. and procedure, substituting a” fer bt® in the stress formula, and a* for bt? in the Deflection formula; where @ = dimension of square. Round Wire Flat Springs Exactly the same formulas (and design procedure) follow in the case of a flat spring made from round wire — Fig. 3 — except that b and t are reptaced by the wire diameter (¢} 6PL . 3 Stress = eS Deflection a Flat Spring Supported at Each End The stress and deflection formulas are modified when a flat spring is supported at both ends — Fig. 4 — and become SPL PL? Deflecti = he ehecton = Tene Design procedure is the same again Stress = aSIMPLE FLAT SPRINGS > Fig. 3 Fig. 4 In the case of a flat spring made from round wire, supported at both ends, bt? in the ‘Stress’ formula is teplaced by d?; and bt® in the “Dellection’ formula is replaced by d* Design of Gontact Springs A contact spring ts simply @ flat spring designed to apply a certain Pressure at a particular point (centact point) along its length — Fig. 5. Ht can be derived from the standard ‘Stress" formule, rewritten as a solution for toad (P} or actual contact pressure produced when deflected. viz br7S: Pp 6L Fig. 5 deflectionFig. 6 400 350 400) avy lint 250] 2 ; E200 io i 2 150 sexphor - Bron 100 \ 50 Z Mickel + Silver Q SIMPLE FLAT SPRINGS ONS Thieknace « inehex 201 925 030HELICAL SPHINGS. where S is the design maximum working stress for the material to be used (ie. use 89 per cent of the limit of proportionality of the spring material from Table t) The simplest way of tackling design is 19 fix suitable values of spring length L and width (b). from which the required material thickness can be calculated. |i this yields a not standard thickness. then the nearest standard thickness tup or down} can be adopted, and the camesponding width re-calculated to provide the required contact pressure. It can also be instructive. having decided on a suitable spring length and width, to calculate the maximum contact pressure available aver a range of thicknesses for different spring materials. This is dene in Fig. 6 for a spring length of 2" and width 4”, and clearly indicates tne superiority of beryNium copper as & Contact spring material Limitations to Flat Spring Calcutations Whilst the design formulas provide atcurate theoretical solutions, actual performance may be modified somewhat by the manner in which the ‘end (or ends) of a flat spring is (are) clamped. In the case of contact springs, performance may be further medified by the fact that such springs are not necessarily simple beam shapes, but may be irregular in width. Calculation applied ta such shapes is tedious. It is best to design the spring on the basis of a ‘mean’ or ‘typica’ width and check the performance by practical experiment. HELICAL SPRINGS 3 Ini the case of helical springs which are either compressed or extended under load, the spring material is stressed In torsion and so the following hasic formulas apply for round wite springs: i BPD Torsianal stess = —— Deftsction ~ Sar XN where P = load D = mean diameter of spring qd = wire diameter G = modulus of rigidity of spring material NN = number of active cois in the spring (see also Fig. 7)HELICAL SPRINGS Fig. F P “dead tun I The stiffness of a helical spring. therefore, ts proportional to che fourth power of the wire diameter, and varies inversely as the cube of the mean diameter. Both d and D thus have a marked effect an spring performance, Using a wire size only one gauge up can appreciably reduce the deflection, and vice verse. Similarly, only a small merease in spring diameter D can considerably increase the deflection: of a small decrease in D can make thespring much stiffar. The fatter effect. particularly, should be borne in mind when making & helical spring by weapping around a mandrel. There will be an inevitabie ‘spring back" resulting in @ spring inner diameter size greater than that of the mandrel. An undersize mandrel is thus required to form 3 spring of required diameter. The degree of undersize can anty ba estimated from experience since it will vary the quality of the Spring material used, and also the coiling technique. The number of active coils is those actually ‘working’ as a spring. Usual practice is to allow 2 of a turn (or 1 complete turn) at each end 12HELICAL SPRINGS Fig. & TENSION SPRING EOS Small offset Plain Gres hook on side a5 Cc Long round end Long square ond hook on centee hack cn gente —{) é esl } tate ee Ose Cogent cor E v ce @® sail Holf Hook on contre Half eye on contre qiq xl Full eye on side Dovble leer Do qd © Eye ond Hogk shown in line dp © Eye and Hook oF right engles 13 | dn Qng aad ground Mot ell} YW heak on sootre OG Coned end 10 hokd iang snival oye Caned eng with | bole ai ® Small aye on centre a © Smell eye on side al) © Eye ond hook shown et right exales I> © wil eye on sidé oad small eye on centreHELICAL SPRINGS in the case of a plain compression spring to produce paratlel ends. Thus. geometrically the spring has a tatal number of coils equal to N + 14 (or N + 2), the number of active coils being calculated for the required deflection performance. Extension springs, on the ather hand. com~ manly have all the coils ‘active’, the ends being made off at right angles to the main coil, e.g. see Fig. 8. Another important parameter is the spring rate (or oad rate). which is simply the load divided by the deflection Spring Rate = deflection _ _&d* 8ND Where the spring is of constant diameter and the coils are evenly pitched, the spring rate is constant. A spring can be given a variable rate by tapering the coil, or using a variabie pitch. Constant rate springs are the more usual, and much easier to work out. Basically, spring design involves calculating the spring diameter and wire size requited to give a safe material stress for the load to be caried. It is then simply a matter of deciding how many coils are required (ie. how many active tums) te give the necessary spring rate or ‘stiffness’ in pounds per insh of movement. This may also be affected by the amount of free movement available for the spring. The same considerations apply to both compression and extension springs, with one difference. Extension springs may be wound with initial tensien, which mm some cases can be as high as 25 per cent of the safe load. To open the coils of the spring this load must be applied, and only the semainder of the load is then available for deflection. This does not modify the spring design formula — merely the value of the applied load effective in producing deflection. Whilst the working formulas are straightforward, spring design is complicated by the fact that three variaLles are involved in the spring geometry — diameter (D). wire diameter (d) and number of active coils (N}, However, only D and d appear in the Stress formula, which is ihe one to start with. So here it is a case of “quesstimating’ one figure and calculating the other on that basis. Design Procedure (i) Either (a) fx a value for D and calculate d for the safe volue of working stress irom = PO 78 4 aHELICAL $PRINGS (Note: From the value of d? so found the corresponding wire diameter can be found from the tables of Appendix B — there is na need to work out the cube root of the answer ta the formuta} or (b) fix the value of ¢ (from an estimated suitable or readily available wire size}, and from this calculate the required value or D from nod? 8p {Note: Ag2in you can look Up d? directly in the tables of Appendix B.) {it) Check that the sizes are practical. For exampte, if the value of D is fixed the calculated value of d may be a non-standard wire size, In this case, recalculate for the nearest standard size to yield an acceptable value of S. This can be avoided by fixing the value of d to start with, but could yield an impractical value for D. (iii) Having arrived at suitable values for D and d. ¢aiculate the number of active turns required for the deflection te be accommodated. o Gd* « deffection SPO? (Note: you can look up values of d° directly in the Appendix tables}. N= That. in fact is all there really is to designing helical compression or extension springs, provided extreme accuracy is not required. Remember to add on 3 or 1 turn to each end for closed end compression springs. More Accurate Working Stress calculstion by the above method assumes that the spring moterial i$ Stressed in pure tarsion. In fact, further stress is added because of the curvature in the wire. Thus the true stress in the material is higher than predicted from simple calculation, viz True stress = KS where K is @ correction factor for wire curvature (nermatly known as the Wahi correction facter) Untortunatety, the value of K dspends on the spring geometry and thus the spring diameter (D) and wire diameter (d) have to be deter- mined before ihe correction factor can be found. Acml , 0615 4c+4 © where ¢ = D/d (which ratio is also known as the spring index) 15ICAL SPRINGS Having determined a suitable size of spring. therefore, the tue stress should be calculated, using the Wahl correction factor calculated as above. If this true stress works out higher than tne maximum permissibie material stress, then the whole spring geometry must be recalculated through. To save a lot of working, values of K are shown graphically against spring index it Fig. 9, and also in Table I! Solid Height of Spring The solid of “closed” length of 2 helical spring follows by multiplying wire diameter (d) by the total number of coils {N + ‘dead’ turn at each end, where applicable). This length may be reduced somewhat by grinding the ‘dead’ turns flat on a closed end spring see Fig. 10 Helicat Springs in Rectangular Wire Section Similar formulas apply, with wire width (6) and thickness (a) reptacing d—Fig. 11. Also additional stress factors are introduced to take into account the additional stresses imparted by hending rectangular section wire inta a helical coil, a POKK, Deflection = NK. Shess = ab “Ge Values of K, and K, are given in Teble Ill. The spring index, for determining the value of K the Wahl correction factor *, is found as follows. For rectangular wite coiled on edge. c = D/a For rectangular wire coiled on flat, ¢ = D/b For non-critical applications the design of helical coil springs wound from rectangular section wise can ignore the corrections 10 stress, by adopting an appreciably lower value of maximum permissible material stress. This will not utilise the full spring potential of the material, but considerably simplifies calculation poeN Stress = a Deflection = Sib 2a?b| | | \ ta Helical coil springs ‘Wohi! HELICAL SPRINGS Fig. 9 soyong Burdydni ow Od Ratio or Spring Indax WWHELICAL SPRINGS. Fig. 10 "dead" torn . around fot dad" turn ground flot Energy Stored in Helical Springs The energy stored in a compression of extension spring can easily be caleulated from P * deflection Energy = ; Fig. 11 —+ + a wits eoited on edge wire coiled on FlatTAPERED HELICAL SPRINGS 4 With the tapered or conical spring, each coil is of different diameter. wes the spring a variable rate. The stress imposed by any load causing deflection is also variable from coi/ to coil. For design purposes it is the maximum stress, which is most important. This will occur tA the largest active coil — Fig, 12 - and the stress is largely tension. PD Max. stress = orp, K where K is the Wahl correction factor (Note: since stress is proportional to spring diameter 0, it follows that the stress in any coil can be calculated by using the appropriate coil diameter; also that the maximum stress occur when D is a maximum, .@. equal to that of the largest coil) Fig. 12 r Dona} (smallest active eal) “dead” tate — — a “deod" tum q Ey fe = Dy af angen serive coil) Deflection under a constant load will vary, In the case of a com- pression spring, first the largest coil will bottom, then the nex largest, and so on - Fig. 13. An extension spring will ‘open’ in a similar pro- gressive manner. 8PD2N Gd* where D, is the mean diameter of the smallest active coil N is the number of active turns Total deflection = 49TAPERED HELICAL SPRINGS 170 Fig. 13 largast active coil bottoms first This formula can also be rewritten in terms of the maximum load to close the spring solid __ Gat * deflection 8D2N The design of tapered springs. therefore. follows the same lines as for helical coil springs (Section 3), using these modified formulas. Solid Height The solid aeight of a tapered spring is less than that of a helical spring dual turns ‘stack’ to a certain extent — Fig, 14. The 20TORSION SPRINGS effective height (y) per coil can be determined from the right-angled tiangle shown, where d? = x2 or y= dF — KF The solid height af the spring then follows as Salid hoight = Ny where N = number of active turns. Remember to add 2d to this to account for one ‘dead’ turn at each end in the case Gf springs with clased ends. TORSION SPRINGS 5 A helical torsion spring is designed to previde an angular deflection af an arm at one end of the spring ~ see Fig. 15 — the other end of the spring being anchored. The stifiness of such a spring (or its resistance 10 deflection is directly proportional ta the fourth power of the wire 7? | anchored end ie diameter; and inversely propenional te its diameter, The coil diameter is commonly fixed (e.g. the spring has te fit over a shaft or spindle): ahd thus choice of different wire sizes will have a considerable effect ‘on spring performance. Fig. 15 oe a Ja dgngetion fdegrens) anTORSION SPRINGS The following basic formulas apply: 32PR *K, nd® “ where K, is the stress correction factor for round wire springs in torsion (see Table 1¥} Pp Anguiar detlection (degrees) = see pee where E = Young's modulus of spring material Stress = Design caloulations are again based on working the spring material within acceptable limits of stress. The force {P) acting on the spring is applied over a radius (R)}, equal to the effective length of the free arm ‘of the spring. Design calculations can proceed as follows’ fi) Knowing the force to be accommodated and the spring arm leverage required (R), use the stress formula (without correction factor K,) to calculate a suitable Wire size’ go = 22PR _ 10-418PR +S 3 where $ is the maximum permissible material stress (ii) Adjust to a standard wire size. if necessary. (iii) Catculate the angular deflection of such a spring. using a specified value of diameter D, from the deflection formula, and ignoring the factor N. This give the deflection per coil. Then simply find out how many coils are needed te produce the required deflection. This stage may. of course, be varied. The load moment PR may be the critical factor— ie. the spring is required to exert (ar resist) a cenain force (P) at a radius R with a specific deflection. In this case, having adopted a specific value for D, the deflection formuia can be used to find a solution for the number of turns required. (iv) Having arrived at a possible spring geometry, recaiculate the true stress as a check, using the correction factor K,,. If necessary, readjust the spring geametry to reduce the stress and recaiculate the spring. If the spring is to be fitted ever s shaft or spindie a check should atso be made that in its tightened position it does not bind on the shaft. 22,TORSION SPRINGS: N Final mean diameter = D = r where N, is the final number of turns when tightened. This is simple enough to work out. A deflection of x degrees is equivalent to x/360 turns. x =N+o Thus N.=N 360 Remember that the final inner diameter of the coil will be equal to the final mean diameter minus d. Torsion Springs in Rectangular Section Wire Exactly the same procedure is invéived, except that ihe basic formulas are modified slightly (see also Fig. 18). _6PR Stress "5 « Ke 2180P RDN Angular deflection (degrees) = aap Square wire section is Simply 4 special case of rectangular wire section where 4 = b_ Fig. 16CLOCK SPRINGS Energy Stored in Torsion Springs This is easily calculated from the deflection and moment. PR x detiection (degrees) 415 The same formula applies to both round and rectangular wire sections. Stored energy = CLOCK SPRINGS 6 A clack spring is a special type of spiral or torsion spring, wound from flat strip. Main interest is in the turning moment of forgue, and the power such a spring can develop, The stress developed in the spring material can be calculated from the spring dimensions in a close wourd and fully released conditions. IF R,, is the radius of @ particular point in the spring in a fully wound condition and R, the radius of the same point in an unwound condition, a close approximation to the stress is given by: 1 a, {see Fig. 17 for notations) fd Bending stress ($,) = ee 7 Fig. 17 fully released Bully wound Point X | Note that it is the material stress in bending (ar torsion) which applies in this case, net the tensile stress (which is lower}. 24CLOCK SPRINGS The ‘Deflection’ formula can be rendered in terms of the mumber of turns (T} the spring can be wound up. 6PRL T= nEtb this can alsa be rewritten in terms of the stress (S,)} LS, — REt (see Fig. 18 for notation) Fig. 18 Deflection ~ aunbar of twins ff] Unwowne wound up fo L = Length of spring This is by far the more convenient farm, but is not strictly correct since it does not allow fer the effect of curvature on stress (see Torsion Springs). The complete formula for number of turns is thus: where K, is the curvature stress factor (see Table ¥) The feagth of spring {L) can be derived from basic geometry. L=a0N where N = number of active coils 2CLOCK SPRINGS In the fully wound condition R, D=2 (s +: but «RL - fy = Nt oR, =Nt+R, Thus D=2 (r. + Mer BR) =2R,+Nt Substituting in the first formuta L=AN(2R, + Nt} These formulas can be used to determine the required spring geometry, with the mechanical output given by Turning moment or torque Q = PR, Hf the applied torque is known, then the number of turns to wind up. the spring aiso follows directly as. _ 6OL xEPY Horsepower Calculation The stored energy in a clock spring can be relessed at various rates, according tO the manner in which the movement is governed or restrained. Note the relationship Sétweén number of turns (T) and stress. No. of turns (7) to produce f stress S in spring material Thus stress (8) = also: Energy per revolution = To determine the energy produced by a clock spring, proceed as follows: (i) Calculate length L from the geometry 26CONSTANT FORCE SPRINGS (ii) Calculate stress produced from the number of turns available 10 wind up {this must not exceed the maximum permissible bending stress of the material). itt} From the stress, calculate the energy per revolution (£,) {fthe energy per revolution {£,) is determined in units of inch- pounds (which will follow using inch units for the spring geometry and stress in Ib#sq.in) €, = (pm Horsepower = 456,000 EX IRM 6 a suitable approximation: =a t* i - orsay Geaaga a PP The time for which the spring witl develop power also foliews as Tyrpm, in minutes -- i.e. the number of turns which can be wound an. divided by the rate of unwinding in zevolutions per minute. The whole series of calculations can, of course. be worked in reverse. That is. starting with a horsepower Gutput requirement and a known value of maximum permissible stress. suitable geometric propartions for the spring can he determined, together with the number of ‘winding’ turns available for the required rate of revolution and number of compiete revolutions Note: as a practical design feature the diameter of the inner coil of the spring. in the fully wound condition, should not be less than 12 times the spring strip thickness. That is, the spring should be wound on an arbor of this minimum size. If wound up to a smaller diameter the spring is likely to suffer from fatigue effects. CONSTANT FORCE SPRINGS 7 Constant force springs are a special type ef fiat strip spring, prestressed to have a uniform tendency to curl along its whole length.” They can be used in twe ways (see also Fig, 19). {i} Rolled onta a bushing to form a constant force extension spring, because the resistance to unralling is the same at any extension. {ii} Reverse-wound around a second drum to provide 8 constant torque spring, or constant torque spring motor. ~ Springs of this type are made by Tensater bimned, Acton Lene, Harlesden. London, NW10. They are known as Tensstor springs ia this county; and “Neg’ster’ springs in America, 2CONSTANT FORCE SPRINGS +» Om, by rolling onta u bushing oF spc0l produces a centiant force extension spring Fig. 19 Prestressed strip by revere winding onto larger drum produces 0 constant terque speing motor Constant force springs of this type have the advantage of being mare compact when reiaxed, compared with helical springs. plus the fact that very long extensions are possible. Either end can be fixed to produce an extension spring as shown diagrammatically in Fig. 20. The fixed free-end configuration, for example. has proved particularly effective for brush springs on electric motors. Fig. 20 lized bobbinCONSTANT FORCE SPRINGS The constant torque or spring motor form is particularly interesting since it offers a performance far superior to an ordinary ciockwork motor, particularly in the length of run possible and the greater mechanical efficiency because of the absence of intercoil friction. Its performance can also be predicted quite accurately. Extension Spring Design Fig. 21 shows the static parameters of a ‘Tensator’ extension spring. Fig. 21 strip width = 6 ot a 25 Dy — —a-4 | - +| The toad to extend can be calculated directly fram the ioad factor for the material (see Table ¥) and the spring width and thickness. P= bt The working extension of the spring (X} will be speci also be determined irom the actual jength of spring strip. > L- 60, where L is the total length of spring. id. but can (Note’ this formula aliaws far 13 dead turns on the coil). The follawing formulas can aise be used to determine D, and D.. D, = YT2THK + TYED,) tt OF In design this should be increased by at least 10 per cent to be on the sate side. to allow for air space between the coils. D, = 1-2 * natural free diameter of spring. as made. 2CONSTANT FORCE SPRINGS Torque Motor Design (see Fig. 22) The torque output (M) available from a constant torque ‘Tensator’ spring motor is given by the formula ObrD, 2 where Q is the load factor (see Tabie vy The hersepower output can be derived from the rate of unwinding. as with clockwork motors (see Section 6} M= Fig. 22 fe —- cane arstence —-»] rte build-up ok of turns — mh fon output rip width ~ b = rm The following formulas will also be useful in design. 6t Optimum value of Bs = a where Sf = bending factor (see Table ¥) M.St Optimum vaiue of t = iste b,+o0 Optimum centre distance = “yo + 301 ‘2. times natural free diameter af spring, 2s made fT 276Lt + DF * where L = tatat length ‘of spring T27BLt + 02 1 (DN + tN?) approx where N = number of working revolutions of D, 30MULTIPLE LEAF SPRINGS & Basically a laminated spring consisting of a number of individual leaves is no different to a sirgle jeaf spring except that the additional leaves increase the effective thickness and thus reduce both deflection and stress in the individual leaves for a given load. Stress calculations are usually based on the assumption of a proportionate laad on each leaf (ke. proportionate to the number af leaves) Fig. 23 averter elliprie vemi-elliptic, semi-elfiptic cantilever ronunber of leoves Three common configurations for multiple leat springs are shown in Fig. 23. The following deflection formulas apply: Half elliptic: | PL? Deflection = Fon Quarter elliptic: Deflestion = APL? Ebt?n Half elliptic Cantilever: Detection = PL? 2EbPn where E = modulus ot elasticity of spring material n= number of leaves aMULTIPLE LEAF SPRINGS The corresponding stress formulas are (the material being stressed in bending as with simple flat springs} Half elliptic: V5PL Stress = a Quarter elliptic: 6PL Su = ress = ae Half elliptic Cantilever: 3PL Stress = ora There are several possible design approaches. If the thickness of each leat (1) is decided, the spring width (b} necessary to produce the required deflection with 2. 3. 4, etc. leaves can be calculated, using the appropriate deflection formula. For example. in the case of a quarter elliptic spring apt? Erin = deflection This will give suitable spring geometry with 2, 3, 4 leaves, etc., from which the most attractive can be selected. This value of b can then be used in the stress formula to check that the maximum permissible material segs ig not exceeded. If so, then an alternative solution must be adopted (e.g. more leaves and smalier width}; or the calculations re-dane starting with a different (higher) value of thickness {1}. Sometimes it is simpier to work directly from the load the spring will carry. which can be arrived at by rewriting the stress formulas: Load capability: Half elliptic Cantilever = where S, is the maximum permissible material stress in bending. A series of alternative spring designs can then be worked out in 32APPENDIX & terms of different values of width (b). thickness (t) and number of ieaves, all of which would be capable of carrying the required load. It is then a matter of calculating the dellection of each of these springs and deciding on the most suitable one. If none give a suitable value tor deflection, then further alternatives must be worked out, Dearing ir mind: spring stiffness increases in direct proportion to spring width {b) and number of leaves (n}; incroases in direct proportion to the cube of the leat thickness. APPENDIX A SPRING TERMINOLOGY {and standard units} Load (P) is the force in pounds (or kilograms) exerted on or by a spring producing or medifying motion, er maintaining a force system in equilibrium. Load is directly proportional to deflection and is limited by the elastic limit of the spring maternal. Defiection is the maximum movement of a spring from its free length of free position to a specified operating position. In the case of helical coil springs, deflection per cail is equal te the total deflection divided by the number of active coils, fate or load rate is equal to foad divided by deflection, and is thus inversely prapertional to the number of active coils in a cail spring. Free length is the true dimensional length of a spring in its unieacad position Solid height is the geometric height (or length) of a coil spring when it is fully compressed. Active coils — the number of coils in a coil spring which deflect under iaad. End turns or part-turns on a compression spring which do not take part in deflection are referred to as ‘dead’ coil Pitch is the spacing or pitch dimension between adjacent active coils in a coil spring. Pitch determines the number of cgils per unit length. Spring tate is also dependent an pitch, being substantially constant if the pitch is constant. aAPPENDIX A Stress is the operating stress on the spring material under working conditions. It is important both to use the right stress value for the material (e.g. depending on whether the spring material is being subject to tension or compression. bending ar torsional loading); and. algo ensure that a maximum permissible stress figure is not exceeded. The latter depends on both load and frequency of deflection. Mean diameter (D) The mean diameter of a helical coil spring is specified as the diameter to the centreline of the coil. The overall diameter of 2 coil spring is thus equal to D + 4: and the inner diameter of a coil spring to D— 4d. Note that diameters can vary with working in the case of a torsion spring. Wire diameter (d) the actual diameter or wire size used in a spring made from round wire. Spring index. This is the ratio D/d and is used to determine stress comection factors where the stess loading on a spring is not simpie (e.g. helical compression and extension springs. and torsion springs) 34APPENDIX B: WIRE SIZES AND VALUES OF d? AND d* 3 jee " a oo | tee co we cae i aoe Bm ; on Gossomy sbinogineesd | eso. pewomeco. canyzmn 139 @qa0cn 428 GQOdINQNDOTIS Qoe2 vy OOuzasas = 9 DUDA FBy 1 qoaood7ia7 g-eaoouaozeset | | 06s coors — eoacorszsz0 | 8 seacooraa pagcooagsas| | |, aoad Denozeza §— q.ag0I67F72 DonoeteaTs GaMGGU SES | tua; Ueqanses — @uuno\ ee ™ neceeecose Snoooaees | Toss Gosia ovootesr? i oooneisrs —deaooeanense: || Dos? o0eg0o7s © owaDazeToN ‘os Sommaoagse seaqgearscgn| foes Geemmag — psmooy Gintoones ooetey| | Gees Goose Simone | sano seacooesse —saonoooasose6 | soo ones socom Som Simonsset Pemeoorseess| | ears eooestar Sussagarer Foe Sole Soootitess tmoeieran! | os Seth Seenaes — Suamedianas i Gust Saaooraszs Demupooaimao| | Sara Uauosiz SumnereneEe Goss ocanoisess | sonoupoeeusze none omenmes — Domusrene | Gate Receaioere Ssamooneseato| i Sort Suonetuae — ommeaaaed Sees SEolaeey angopoes cup Boer Goyeasesy — Siwogersog spa oacoNz1SS2 daaoeNAetSses| D039 Docdozare9 © o00000r07780 : i : n¢30 geqcazro0 © aaaqgus1900 Bear pegoozars) ogagaaagess? j 2 paz Seacone76s GnoocoT OMNES 0053 waCORaSEs7 © HetGOgTIaRET Ouss aoncozsaaa 4.000001 sans t Dass AaMOMZETS MEMO TSIOET | 20 eas a coonences on eagonteraa2 i 0037 Daponsdcs3 © ueagaata7s16 0038 v00nsCB72 NNGCQUzURETE i OnGH GpCDOFAIA —a-otoenaaTTA4 OTS GnoMKTI oem TI, 35 “wopccEs0e — o-HOORIARODT 1s 0m ees awecouosion | tear Soquates — osenagee? Soe Saiuerigs Bamaetesss Does Semayes Sb omeTies Gos Snonazrs — bosearen Soes Senmersie pomearert ose Somcesnve op aamestene oso Spuensa ft pos aeye Sose Srestis < gomessese Sass De vesse Boomer 12 9G4g a pCOEEaDDO BAN ADETIN i 0081 @.00pe82) 9 onoDOREDS 76 ase vemanz608s § — osaages14 170 9003 9 9000?807 goNoDNUaTEAO Onda GOOONESTa4 — G onDDeNT4RIO I Oo0§ a D00081 125 a pLOOWAICNGA ORK a-q0o087335 4 on0DNsTT 748 9087 G.g00nNEZs —o noDODABT BAR | 1&8 OAR GoONTTEES? a MnOnAESDBA? 9.959 aege7zse0 —o s0ncss6100 Gost tamiaes — oooemnaua 1s Ouse paaozvees — Gououriesss 1! Soc Geoosnere — o-coa0 70857 ooe4 atone © ocokeroTas | Sos aeooesnss — pocatst sto cee Gime: gomanee | Bow Somsizer — Donenasszes Goa, Gitceuirs §— Gomaceraea | Obes coonritede 0 aoONETCEED Goes Oonosvesn 0 sounscosne | I 0080 eager z600 — @ anoovEzENC0 | Cast Gaooviane a banoos2ese Oost gogctaest — omoenrants Coss Nopovasa — Roscaardads Sos Oocolsras a oopaesat Goss nooders §— ceqosnatses 17 fase Goootraee © SonoomeS 00s, ooedeta — OonontaceeD Sosa “orsent — Souor saa 059) 0 000438 atoboT «| 0100 semen a eoerneee orot Gomigdso © SoDbINDE Gree damaaige — gaootagee | Sica bhowbers © Gaeoltess Isa Gta Daortzese — oooctiess | ois snore = ooooraiss Oto, Goptietoz — Gonelze2 | O12 Deowztsoe — o-oama1a8 Otos banner GoogT eos Saga aoe OOo ATT 35APPENDIX B: WIRE SIZES AND VALUES OF d? AND d+ r 3 1 Pa j a won ro * op wm o « G48 poorsste oc ouoraea: 2 b160 gonsso oaporsaae Ont eotiers — GopeTstat Sie suns © Samer | oi Banas Badotsrse | len Soczsis — Sacoeas7s Gira oes Sooutezoe e109 Goossen S.e00;aest Sia amass Beoersose | Grea Goccares — Oogorass Cals Geniszos © Gace Bier Ghown — Saverante uote Geiss © couotator oie aeoteras © Gaca7ssio ov? Baowns = Sodoierss | Ges Sooesrs — oougrymeD Gre goer — Saoniaiee Sieg feet = Soograess | Gre Gaoteane§— Sacazeaes S169 foeeeree — SonuRteTS | O12 BOO T2580 Dacoz0736 O170 GC onESTIO 9.q0083521 | oan oaa7rs © eongzT4a5 1 173 Geuscenz = QUORBSGM | Bia Sootmas — Saenzztas | Giz Saanones — Sooogress Sia oooweees — oanozzess j 80 omens S agcgacts baz Guoieess §— Bowczsnaz Sine Gcsueee — G acosi soe Stas Ganesnst oozes Oi) Geosases ——@ dosazess Oak Gono Soouzseos 7 Orme Geese g oocezast O42 boggerge a onnzedta aire peoseere 8 n00580m1 (10 tame aoeware — Gobotenes eye povesrsa a apron O1m OG oeztaay — Baenzreee 178 nosnds3 © ootnzees 0398 sasava70—ganaagean aim come amos | 01s Bocstea — SpRUZSHSD | seh beasszer — Gactorsie Otsz ponereod — ososenes | Stee bvennee | O-poibense 2333 Gonueze © Gonatasg 1 Gies ooestzes — gottzise Oia oamienh — Gongs33e3 Sad Geusnzss §— Seotnaezs | 0138 ‘Sapiens encaazia oie coeaits ponaas Gia Sandie Genazio Oa86 Gooey Ogoniaeee Gis? Seams besser Cie? Geoecta a o1zza3 C133 ooosezas — Gongere? See Steccudy§—aptzaaz0 13a Sboeenss Senco 185 Geese's§—awonz7ee oensaniis ors novegeso — pger309 oeeneses cist tees: eoisaa B12 GO07BESS— G.b0040NS 9 & 8162 geqroTa © -przs80 B14 pngRg? = apcoatEia O18 SoBe 01385, $ 01a anazaeeo §=—a.po04z 098 C194 GONNA = aDOTHIaS 9145 0030496 = .90040705 0195 aoorai4s == gap naasG e148 gnogtzt «= agieesta? BiS8 GOERS §=— gota 7a 147 DO031765 0-008 0197 © O0oTEMs4 a DOTSOgT Q1eg poms §— 00047979 Diese «9 go7Eze = agNE 70 G49 nomsers = a pnas7ae 5189 agreece == -00 rs 9180 Dooss7s00 nN agEOEZE i 0200 ea0ena = ac 1800 181 DoosaazaR — g.onoRseay 9182 Gooastiay §— 0.60053379 F 07IR onogsae 9002078 i gis geniseteg © goudga7ag = || «ots? ousies = ounansr D156 O80R8Sr¢0 © D GnUEEDES 1s5 0.00872390 0 00067725 3) D252 on90a Ge OANOS Grek eoonieeie — pecasa22 7 O86 GorteRe =e 0osea. 157 ©.00986990 0. Q0060757 ois3 Guussease © Sopoaeany + 900 ensreen =p. uesta0 O1ks 9.00MDND70 —0.90083913, 10 0934 como = nortan - u.Let ot Menimar safe warineavess frie ws af Manes snapavtinaboa is ‘- ‘Sant ule ea nt vexsiow went Eten ont Pinar upra@1 fa iam suenanoe ‘Onl eembered sie sae 190.000 aroac00 arg aroun etl ae 190 990 Porc Stamtass sige’ 18 8 we ” we eomO = Hannon Sramtess ate very - ‘30 600 ranoa.pan Etvoeme vanaduer ‘enn a0 Prospact teonne” rasan se000 60.100 Tsoon300 Bue S200 35000 rooney Nel sam 50 080 zane O Servite copper” 188 118800 = 16+18.500000 PLenel sor" 5 6a000 - raooa oe F Banas «pang ws ~ . - - Ue Be percent at hee rue lon aera cal ROnEAC: ScHreH TABLE Wf WAHLS COR- TAGLE Ill RECTION FACTOR K FOR ROUND WIRE MELICA COIL Cerne 2.000.000 N50 090 1/500 600 AKANO nso Gp v'st0 en" 6300 coy Seonees S600 600 6 7ean000 ssmcap CORRECTION FACTORS FOR RECTANGULAR WIRE HELICAL COM SPRINGS: SPRINGS | Number stress Deflection Spring mdex Dit x of active facior factor cots Ka K | 2 2-06 1-0 480 555 3 158 19 426 aor | | a 140 29 290 3s 5 ey 28 372 307? 8 135 30 360 292 | 8 118 40 348 276 i ; 10 444 80 3-30 ae} i 12 112 10-0 345 250 3 i 15 08 20 3:08 2a} | 20 1.07 50 a0 a | ! 25 106 100 3-02 237 TABLE IV STRESS CORRECTION TABLE V DESIGN VALUES FOR FACTORS FOR TORSION SPRINGS “TENGATOR” SPRINGS f = oat of active round rectangular | ff@Mo.of steel steel coils wiraK, wins K, i cyoles QO S GO %& 2 161 15a e000 21 0023 680 O027 3 133 129 10000 418 0020 S02 9-023 4 123 1.20 20.000 2 OMS 350 0013 8 118 118 4.000 169 0-010 233 0-012 8 aaa 413 70.000 125 $003 154 0-008 & 110 109 200.000 101 G08 «87 0-008 10 1-08 107 000,000 «Bt 70 2 1-08 4:06 % 105 1-04 20 1-08 1-03 25 104 1.03 30 1-03 1-02 above 30 10 10 7