8/8/2019 1979030

1/6

A new heat-treatment process overcomestemperature relaxation problems forspring usersR. G. Slingsby

High-performance springs are demanding increasingly greater use of the potential available from the correctheat treatment of modern spring materials and application of current manufacturing technology. This is parti-cularly true in the design of modern internal combustion engines where load loss o frh e valve springs when operat-ing at running temperature wifl cause serious problems. Spring relaxation can be reduced to acceptable levels bya process known as 'h ot prestressing' or 'hot scragging'. This paper discusses the mechanical and thermal pro-cesses involved in the production of high-performance springs and shows how it ha s been necessary to study theinteraction of the variables in order to obtain an optimum com promise between atigue life and stress-temperaturerelaxation. A new concept o f resonant frequency fatigue testing w as developed and a brief description of thedesign of the rig and control gear is included. The development of special-purpose machines to carry out hotpresrressi~gon a large production scale is the logical conchisiou to the work .The autho r is with S alter Springs Ltd, Smethwick

AUTOMOBILE VALVE SPRING MATERIALSThere are two major categories of the h~gh-performancewires from which the springs are coiled: ( a )patented andcold-drawn carbon steel ; (b )carbon and low-alloy steelswhich have been quenched and tempered subsequent todrawing to size. This latter type is usually termed 'pre-hardened and tempered wire'. Typical compositions aregiven in Table 1.Both types of materials are produced to a tight speci-fication for mechanical properties, freedom from de-carburization and surface defects.The low-alloy prehardened and tempered wires ar ebecoming the preferred material since with suitableprocessing they offer the best combination of highfatigue performance with low relaxation.

SPRING DESIGNThe actual design of a valve spring is a very complexmathematical exercise requiring an analysis of the camform and the mass of the moving co mpo nents of the valvetrain. The final specification will define the dimensionsand the loads. A typical spring is defined in Fig. 1.MANUFACTURING STAGES 'Automatic coiling machines are set up to coil the wireinto the required form. Th e shape of the 'as-coiled' spring

will allow for changes in geometry which will occurduring later stages in man ufacture du e to the removal ofcertain stresses and the deliberate introduction of others.The major stages are listed below (not always in thisorder):(i) coil(ii) first low -tempe rature stress relief(iii) end grinding(iv) shot peening(v) second low-temperature stress relief(vi) cold prestress(vii) hot prestress.

Most of these operations will influence the fatigue andrelaxation characteristics of the final product and theseare discussed in the following sections of this paper.EFFECT OF THE FIR ST LOW-TEMPERATURESTRESS RELIEFT h e purpo se of the initial thermal treatment is to removeundesirable coiling stresses and to obtain an improve-ment in the elastic properties of the wire.Some of the changes which take place a s the tempera-ture of treatment for cold-drawn wire is varied areillustrated by Fig. 2. It will be seen that although themaximum improvement in tensile strength and elastic

8/8/2019 1979030

2/6

Overcoming temperature relaxat ion problems for sprlng users 11 7Table 1 Typical chemical compositions of valve spring wire

C ,% Si ,:; M n , % S, ? < P, 9 : Cr, V,O0Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max.

Patented cold-drawn 0.55 0.85 - 0.35 0.30 1.00 - 0.030 - 0.030 - - - -BS 5216, code HDPrehardenedandtempered 0.47 0.53 0.15 0.30 0.70 0.90 - 0.020 - 0.0250.80 1.10 0.15 0.25chromium-vanadium,AVOT 2Prehardenedandtempered 0.60 0.75 0.15 0.30 0.50 0.80 - 0.026 - 0.0300.30 0.55 0.05 0.20chromium-vanadium,AVOT 3

limit is in the order of 200"-250C the best fa t igue andstress temperature re laxat ion occur a t abo ut 350C. I t i scurrent pract ice to use the higher temperature andaccept the fact that t h e op t imu m sol id s t ress is notobtained.The chang es wi th t r ea tmen t t empera tu re fo r p re -hardened and tempered chrome -vanadium spring wireare shown by Fig. 3. In the case of th is type of mater ia l , i tis found that once coiling stress is relieved at temp eratu resabove about 350C increasing temperature will causelittle change in fatigue performance; an improvement instress temperature relaxation is however observed as thetreatment tempera ture is increased up t o a level approach -t so110 loa d 1F2 va lve o p e n I

L , , ,'-ec MOVEMENT so t id,er;?'- l e n g t h4

f r e e i e n g t n7- - ----1 Definition of t v ~i c a l alve sorina

ing the or ignal tempering temperature , which wil l be400"-450C depending on wire size an d tensile strength.It is usual therefore to stress relieve after coiling atabou t 350C for cold-drawn and a t 400'-450'C fo rprehardened and tempered wires.S H O T P E E N I N GThis process consis ts of bombarding the surface of thewire with spherical hard steel shot to produce a com-pressive stress in the surface layers of the wire. It is ofpar t icular im portance to peen the ins ide diameter of thespring because this is where the greatest srress is deve-loped when th e coils are compressed. T he size of shot, thespeed of the impellor, and the timeibulk ratio are allimportant factors and are the subject of cont inuingresearch. Two types of shot are in current use: ( a )hardened and tempered cast s teel ; (b ) 'cut wire', which ishard-drawn high-tensile wire, cut to a length equal to itsdiameter and then processed to knock off a l l the sharpedges to produc e a roughly spherical form . The la t ter ispreferred by Sal ter Springs Ltd , the advantages being aconsis tent s ize and hardness , no breakdown, and long-00 250 300 35 0 4 0T EMPER AT U R E, 'C2 Effect of low-temperature heat treatment on pro-perties of cold-drawn va lve spring wire

8/8/2019 1979030

3/6

118 Sl ingsby

3 Effect of low-temperatur e heat treatment on relaxa-tion of prehardened and tempered chrome-vanadium valve spring wire

service life. (A fractured shot particle can easily cause anotch in the wire surface with disastrous results.)Shot peening elevates the fatigue limit and permitsdesign with a much larger stress range than that obtainedfrom un-peened springs. The results of fatigue research

I I I I0 200 400 600 800 1000IN IT I AL TO RS ION AL STRES S, N / rnm' (co r rec:ed)

4 Goodman diagram for a cold-drawn valve springwire ; b Avot 2 shot peened

5 Resonant frequency test rig and control unitare normally presented in the form of modif ied Goo dma ndiagrams, Figs. 4a and 4b .Contro ls of shot-pee ning processTh e usual technique of measuring th e effect of this processis to specify an 'Almen Arc Rise' . One or more hardenedand tempered steel strip specimens are mounted on ahardened block and put through the process with thebatch of work. Peening one side of the specimen onlyputs an unequal distribution of stress into the testspecimen, which curves due to compressive stress on theconvex face. The a mou nt of curva ture is a m easure of theintensity and coverage of shot peening.We have always been dissatisfied with a test specimenwhich bears no similarity to the components which it issupposed to represent. Attempts to relate Almen resultsto fatigue tests on actual springs have not been veryconvincing which is not too surprising considering theway in which a heavy block may randomly orientatewhen tumbled with a charge of several thousand springs.Moreover a spring is free to react dimensionally as thepeening stress is introduced w hereas the Almen specimenis restrained all the time.A test method was therefore developed to producequick fatigue results from test samples which are verysimilar to the springs being processed.The testpieces are lengths of wire coiled into open-ended springs of regular pitch. Having been through allthe process stages with a batch of springs. these testpiecesare mounted into a rig and vibrating energy of lowforcing signal power is applied t o one end of the coils.If the frequency of this vibration is adjusted to matchthe natural f requency of the tes t sample , a h igh ampli tudecan be g enerated in the free coils with a resultant stress

range of sufficient intensity to cause early fatigue failure.In practice, it is difficult to obtain a co nstant resonancecondit ion due to such changes as ambient temperatureand slight shift in the bearing points. This was overcomeby having a ma nua l control for initial set-up and thenmonitor ing the ampli tude of the spr ing coi ls us ing amagnet ic t ransducer to feed a s ignal in to an automaticcontrol system.' (De veloped by the Dep artm entof Electrical Engineering, University of Aston inBirmingham, Fig . 5 .) The forcing signal frequency andampli tude are then adjusted cont inuously to maintainresonance. The system also terminates the test when thegrowth of a fatigue crack has altered resonance by a pre-determined percentage. Thus all tests are terminated at avery similar stage of crack propagation and the scatter

8/8/2019 1979030

4/6

Ooercoming temperature relaxat ion problems for sp ring users 119

resonant f requency techniq ue

5.6

54-

..-.~nu-x5 .2 -u

2;=5.0-S

4.6

normally experienced with more conventional springfatigue tests is reduced.The technique provides a quick method of com-parisons. For example, for shot-peening control againstpredetermined standard s, three test springs are employedand a control chart is obtained from the average andrange of the test results. Th e effect of varying processtreatments can be established and the characteristics ofdifferent types of materials can be explored as a pre.liminary to conventional fatigue testing..SECOND LOW-TEMPERATURE HEA TT R E A T M E N T

'' 225 250 275 300 wards the neutral axis being progressively lower is belowTEMPERATURE, "C the elastic limit. This is done by manufacturing the6 Effect of post-shot-peening low-tempe rature heat spring with a free length such that the theoretical elastict reatment On fat igue l i fe of Avot 3, obta ined f rom limit of the spring materia] is exceeded when the maxi-

The shot-peening process will produce what are common-ly termed 'micro and macro' stresses dependmg upon theang le o f rnc~dence t wh~ ch he sho t par t~ cle on tacts thesurface of the wire The pure com presswe mac ro stress ISbeneficial but the micro stress may reduce the elasticirmir A fu r ih r r l o w - rem p era~ u r r h rd i i r rd ~ m en t w ~ i lremove the undesirable com ponen t of the induced stresses

but if carried out at an excessive tem pera ture will removethe beneficial stress and so reduce the fatigue life of thesprings.It has also been found that relaxation is increased byshot peening but reduced again as the temperature of thesecond stress-relieving treatment is increased.The effect of the temperatu re of this treatment on sho t-peened springs had been studied using the resonantfrequency technique (see Fig. 6 ) . This indicates that theoptimum temperature for chrome-vanadium pre-hardened and tempered wire is in the orde r of250C.COLD PRESTRESSINGThis process increases the apparent elastic limit of thespring material by introducing a stress pattern in whichthe elastic limit of the outer fibres is exceeded so thatplastic deformation occurs, while the stress further to-

I I

T -

7 Simplif ied mechanism of prestressing9

I I

mum possible compression is applied (solid height).When this load is relaxed, the fibres of the core attempt torecover their original position but are restrained by theplastically deformed outer fibres: these o uter fibres, beingstrained by the recovery stress of those inside. only partlyrecover their original position and so a condition ofequilibrium exists in which the resultant stress patternallows a solid height which is greater than that which istheoretically possible. Therefore the load capacity of thespring is improved and the o ptimum use of the material isobtained. In practice the process of prestressing (or'scragging' as jt is known in the trade) has to be carriedout three times or more to obtain a stable condition(Fig. 7). After prestressing, springs may show slightchange in load over a period of time, or if 'bounced', andthis probably accounts for some of the variations inload between different inspections.The extent to which a spring is prestressed will bedetermined by the solid stress needed to n c r t the loadrequirements and the mechanical properties of the springwire, which vzry according to the wire diameter for hothcold-drawn and prehardened and tempered materials. Insome cases the physical proportions of the particularspring design inhibit the amou nt of prestress because the

spring will distort if closed solid.This cold prestressing operation has little influence onthe relaxation of springs which need to be finally hot pre-stressed but will often be carried out possibly with onlyone closing, or 'scrag', in order to make the spring morestable and to reduce the total time taken for hot setting,particularly when the spring is heated by a resistancetechnique discussed later in this paper.HOT PRESETTINGThis is the operatron whlch needs to be added to normalspring manufacture !fit becomes know n tha t the relaxa-tlon of an o ther w~s e esirable sprmg design IS more thanthat which is acceptable There are several combina tionsof the methods wh~chmay be employed to hot prestress acompression sprrng. The prac t~ca bih ty and technicalaspects of these may be considered as follows below

n o r mm o d ~ a n-

Method (a)This consists of clamping the spring either solid or to apredetermined length and then heating spring and jig for

-, -

,,qn/-'

.. -.L

-.8 -med ian 1' ,/- 3 6 ,norm /- 3 u . T ," A I

8/8/2019 1979030

5/6

120 Slingsby

na x strass 1100 ~ l r nm ' 1; max. strQss 1 2 5 0 ~ 1 r n r n ~ -$ 1 , , , , , , I0150 200 250TEMPERATURE, 'C

8 Effect of hot prestressing temperature and stress onrelaxation of shot peened Avot 2

sufficient time to allow the major part of the 'creep' totake place. At the conclusion of this, the spring can becooled to am bient tem perature while in the jig, orquenched. Alternatively, the spring can be ejected fromthe jig and cooled in air or quenched.In this technique, the best results are obtained if thespring is quenched while under restraint so that themaximum amount of prestressing is obtained. If thesprings are removed f rom the jig when hoi, a proporti onof the induced residual stress will be dissipated by thesimple stress-relieving effect of the heat contained in thespring.The economics of this method are not attractivebecause of the need to employ large numbers of heavyjigs which are continually being heated and cooled. Thejig is likely to weigh more than the spring. There mayalso be problems in making the jigs adaptable fordiffering designs.

Met hod (b )This consists of heating the spring for a particular timeand temperature, then closing to solid or predeterminedlength, followed by quenching while under restraint. Thistechnique is likely to be more economic if the springs arecontinuously fed thro ugh an efficiently designed heatingsystem. Th e transfer an d closing mechanisms d o not poseany serious design problems.The springs, however, are cooling as soon as the trans-fer and clamping have been completed and the resultanteffect on relaxation will not be as effective as methods inwhich the springs are under restraint while the heat isbeing introduced .This method is being used on a commercial scale and inmany cases is producing acceptable relaxation values.M e t h o d (c)The most attractive method, from the economic andtechnical points of view, is one in which just enoughenergy is used to bring the spring up to temperaturewhile it is compressed to a level of stress such that thedesired amount of creep is allowed to occur and theprocess is cut off at a particular point for individual

springs so that the resultant loads at length are veryconsistent.Accordingly, experiments were carried out to examinethe feasibility of using low-voltage a.c. to heat the springby resistance and then to devise a system which wouldstop the process at the necessary point.T h e results indicated that satisfactory relaxation couldbe obtained together with a close load tolerance (of theorder of 3 4 %), often better than normal 'cold scragged'spring production. It was also found that provided theelectrical contact plates at each end of the spring arecorrectly designed and maintained and a critical force isapplied before the current is passed, consistent heatingand freedom from 'arcing' can be established. 'Sparkguards' can be integrated into the circuits as an addedsafety factor. The applied current can be adjusted for aparticular design of spring to ensure that the temperatureof the spring does not exceed that at which the fatigue

performanc e would be adversely affected. (The springs areshot peened when this operation is carried out.)Finally, a system of mechanical handling and auto-matic sequencing was developed for which patents arenow pending.The equipment which was developed has progressedthrou gh several stages of refinement and is now complete-ly reliable on a high-volume basis. The economics aresatisfactory and sample checks are regularly carried outto prove the fatigue and relaxation properties of springsunder d ynamic stress. Fu rther confidence in the techniquehas been obtained by extensive road testing and by theperformance of standard production vehicles which arenow being regulariy fitted with valve springs which havebeen hot prestressed by these machines.Effect of hot prestressingThe extent t o which relaxation of a spring under elevatedtemperature can be reduced is mainly controlled by twofactors. The first is the stress level at which the hot pre-stressing is terminated. If this stress is increased therewill be a red uction in relaxation. The second factor is thetemperature of hot prestressing. If unpeened springs areconsidered first. there is a reduction in relaxation as thetemperature of hot prestressing is increased to about400C for prehardened and tempered spring wire andabout 350C for cold-drawn spring wire. If the springshave been shot peened the temperature of hot pre-stressing is limited by the effect this treatment will haveon the fatigue limit. as previously discussed. However areasonable compromise can be obtained if the tempera-ture is not allowed t o exceed 25OoC,Fig. 8. The relaxationshown by these curves was obtained by static tests. Ifthese conditions are repeated under dynamic conditionssimulating actual valve gear operations it will be foundtha t the relaxation will be considerably reduced. In someinstances it will be only half of the static relaxation.CONCLUSIONSThis pap er has at tem pted t o show that a relatively simpleheat t reatm ent such as stress relieving, combined with themechanical processes of spring making and other pro-cesses such as shot peening, can be studied to optimizethem for performance under high fatigue stress andelevated temperature conditions.Th e use of the hot presetting technique is by no meansconfined to automotive valve springs. There are manyother applications for the use of heat-stabilized springs

8/8/2019 1979030

6/6

Overcoming temperature relaxat ion problems for s pring users 121made from low-cost materials. T he technique is applica bleto som e of the more highly alloyed spring materials and itis anticipated that the application of hot presetting tosome of the problem areas in which exotic alloys are atpresent used may allow the use of lower alloys withconsiderable cost saving and possibly better all-roundperformance.ACKNOWLEDGMENTSThe author wishes to thank the Spring Research andManufacturers Association for permission t o publish the

I4

Goodman diagram for hard-drawn valve spring qualitywire abstracted from their publication 'Springs, materials,design manufacture'.He also wishes to thank Professor Barker, of theDepartment of Electrical and Electronic Engineering,University of Aston in Birmingham, for permission topublish a photograph of the resonant frequency springfatigue test rig and its control unit.REFERENCE1 H. A. BARKER ef af.: roi;lnrt. Electr. Eng., 1978, 125, (9)