690055 New Criteria for Self-Loosening of Fasteners Under Vibration Gerhard H. Junker European Research and Engineering, Standard Pressed Steel Co., Unbrako Koblenz Reprinted March 1971 from 1969 SAE Transactions, Vol. 78 by SOCIETY OF AUTOMOTIVE ENGINEERS, INC., ‘Two Pennsylvania Plaza, New York, N.Y. 10001New Cri 690055 iteria for Self-Loosening of Fasteners Under Vibration A UNIQUE TEST method and apparatus now make it possible to reproduce conditions of vibration that are not only certain. 1 loosen bolted joints but which also closely simulate actual ‘conditions. Ir can be shown that properly prelosded fasten- fers loosen as 2 restit of rotation as soon as relative motion ‘occurs between the mating threads and between the bearing ‘surfaces of the fastener and the clamped material. In addition to fatigue failuee, self-loosening ts the most frequent cause of failure of dynamically loaded, bolted joints. ‘This may be the reason for che voluminous literature on the operation and effectiveness of self-locking elements de~ signed to prevent self-looscning of threaded fasteners, with resultantsevere damage tomachine and vehicle assemblies. ‘Most of the existing Mterature, however, gives she impression iat there have been many approaches tied without finding, the correct solution to this problem. To be sure, much test data have been abtained in both the United States and Europe in an attempt to explain these action-feaction phenomena. ‘Unfortunately, up t0 the present time, there does not seem ‘to have been sufficfent exploration to develop any trily practical answers. Goodier and Sweeney (1),* as well as Sauer, Lemmon, and Lynn (2), have tested only axial dynamically loaded, bolted connections. Ia spite of their failure to obtain @ Numbers in parentheses designate References at end of paper. ABSTRACT Gerhard H. Junker European Research and Engineering, Standard Pressed Steel Co., Unbrako Koblenz complete self-loosening of threaded fasteners, they have ‘offered an explanation for partial loosening, which does contain the basic element of the anechanism of self-loosex {ng of bolted connections, That is, bolted connections are oosened by the relative motion between the thread flanks and other contact surfaces of the clamped and clamping, parts In 1966, Paland (2) precisely formulated and affirmed the prior theory and also presented exact proof. Nevertheless, neither could he produce total self-loosening (that is, the total loss of clamping force), even by extreme axial impact loading. ‘ie 1966 the author approached the problem fom another direction. My peper, "Investigations of the Mechanism of ‘Self-Loosening and Optimal Locking of Bolted Connections” (4-8). presented data from tests mainly carried out on trans~ versely loaded jofnts. There was no difficulty In completely rotating loose noalocking fasteners as well as some so-called self-iooking fasteners. ‘New tests, which affirmed that transverse vibration gen~ erates the most severe conditions for self-loosening, ed 1 the design of a new testing machine, which is supposed to _pive practical answers to the phenomena of self-loosening fas well as quantitative data on the locking performance of self-locking elements. After a general discussion here of the mechanism of self-loosening. this new machine end its operation will be described. The various possibftities of recording and evaluation will be shown and discussed, with, ‘Tho theory of seif-loosentng of preloaded bolted comnec~ tions when subjected to vibration is discussed. ‘The signifl~ cance of self-loosening as a cause of failure is explained, ‘and design guidance t avold self-loosening fs given. The Test methods are described and discussed in connection with a4 ‘a newly designed testing machine that yields quantitative ata for evaluating locking properties. These methods can be applied to all types of locking elements. Finally, « sim- plified method for broad scale vesting and inspection fs pro~ posed. £} weon the aim of proposing solentific and sophisticated test methods torque fs applied wile pulling. If car fs braked in a curve. 1 well a8 a simplified inspection test. the ear lose its sideway geip on the road. ‘The example of It isnot the intention of dhis author to compare the per the slope is also applicable to a bolted connections The formance of self-locking elements, ot even of those that thread is the slope, spirally mounted, and the clarnping load have been wed in these test series for the purpose of evalua~ is the weight that causes the pressure between the two solids tion and tial of new test methods and machinery. To give in contact. Additional frletfon forces originate from the final answers on the self-locking performance of various clamping load oa the bearing surface of the bolt head ot elements, much more test data must be obtained on a sta~ the aut ig. 8) Ustical basis after agreement on suitable test mediods has Similar to 2 load lying on a fixed slope, self-locking been reached. exists in a bolted connection as long as no relative motion arlserbetween the thtesd flanks and the contact surfaces of ‘THE MECHANISM OF SELE-LOOSENING clemping and clamped parts. The following off-torque fs racer, en ee [ocho Sa eel alee ela gua CO eee cca eee cra bolt head or nut bearing surface. The resulting transverse 4, D ] ee | ance + py + le tance + Oy ‘The bolt could loosen by itself (without the application of extemal off-torque) only if T. were to become either Q= Liane +p) +Ltanp wo ‘As long as Q is greater than zero, the system does not move. It will move, however, as soon as the slope underneath the solid body is vibrated to the extent that the inertial force created exceeds the friction force so that the interface be~ tween the solid body and the slope becomes apparently free of fefotion. As Fig. 2 shows, the transverse load L tan @ then injects motion into the system. Oe eee This effect can be demonstrated by examples from dally ene Lifer A cork can be more easily removed from a bottle if Fig. 2 ~ Simplified model of a bolted joint under vibration Qs L- tan (-9 +2) +L-tane Jott #Fe[SF ten (9 0) + 204] Fig. 1 ~ Simplified model of a bolted joint under statie ccondisions Fig. 8 ~ Bolted joint under static conditions 318316 a zero or negative value. Even under {deally lubricsted conditions, the thread friction angle will not be less than 6 deg and the helix angle will not exceed 3°20" on sizes down to MS diameter (equivalent to 9/16 in.). Therefore, the sum between the square backets in Eq. 2can never be~ come smaller than, ot equal to, zero; that 1s, the prod- uct of the equation can become zero only if the preload drops to zero. ‘This ease, which exists only if no relative motion oceurs in the threads or at the beating area inter faces, has been very widely generalized. However, if relative motion occurs between the threaded susfaces and/or other contact surfaces of the clamped and clamping parts because of an extetnal force, the direction of which is either tangential or radial, the bolted connection ‘will bocome free of ftiction in a circumferential direction Gig. 4). ‘This means that the preload acting on the thread, which i a slope, ereates 4 force in a efreumferenttal direc~ ‘Hon and results in the rotational loosening of the bolt or the nut, The maximum value for total elimination of the ciccumfereatial fiction force resisting the existing internal off-torque is then equal to 1, tan -? = @ ‘The question now is: How do these relative motions acting between the thread flanks and other contact surfaces of the clamped and clamping parts occur? For axially loaded joints, Geodier and Sweeney have already pointed out that pulsating tension of a clamped bolted connection creates radial sliding motions between the thread flanks of the bolt and nut or at the interface of the clamped Dearing surfaces. The reasons for this are the contraction of the bolt according to Poisson's ratio and the dilation of the aut walls caused by axial tension. Paland proves this rule arithmerfeally and by measuring the tangential strain. fon the surface of the nut. He comes wo the conclusion that ‘a loaded nut widens elastically in a radial direction at the Fig. 4 ~ Bolbed joint under transverse vibratfon GERHARD H. JUNKER n area neat the bearing surfaces and contracts in the upper pact. For the widening area, for nuts of M10 diameter (ap- proximately $/8 in.), he specifies values for the relative motion as Sf = (1-++8) 10 mmAge @ “These very small amounts of radial displacement by ex= pansion ofthe rut would explain why Paland (3), fa spite of heavy impact loading in an axial direction ofthe bolt, suit needed ¢ small extemal off-mrgue t9 tun the nit com pletely jose. Goodier and Sweeney (1) have succeeded at ‘the end of 500 cycles in producing angles of 6.5. x 107° revolutions, approximately equal 10 2 deg, when loading 4 9/4 in, nut-bolt connection on a tensile testing machine. Nevertheless, this loorening had already caused a preload sop of 21%. Sauer, Lemmon, and Lynn (8) loaded 2 5/16 fn. uut-bole connection ia a fatigue vesting machine with « working load quel t0 80% of the preload. ‘They recorded « maximum loosening angle of 6 deg after 25,000 cycles when using cadmium plated nuts and bolts. Their loosening curves chow that the rotating process had come to a stop. The recorded rotation loosening angle of 6 deg nevertheless had caused a decrease of the clamped length of 24 bm: tis is equal to a decrease of 700 kgf (1550 1b), when related toa clamp~ ing length of 2a, . Using a Grade 8.8 bolt (equal w Grade 5) this reduced the proload by 42%. For dynamically mausverse loaded joints, the relative motion between the thread flanks and the contact surface of the bearing areas can occur in magniudes up to the maxi- mum allowance of the thread. These large effects appear when transverse loadings, which have w be transmitted by azip frfetion, exceed the frfetion force between the clamped atts, HE, the friction force being delivered by the clamping, force F. The resultent transverse slippage between the clamped parts forces the bolt to assume a pendulum move ‘ment, which leads 1 relative motion fn the thread hole and thus between the thread flanks. If the amplitude of such transverse slippage of the bolt ‘is large enough, slippage of the nut or boit head bearing sueface will finally occur and make the joint totally free of frfetion in a circumferential direction. It can be easily realized too that, contrary to the conditions of axial load~ ings, relative motion between the flanks will occur in ail parts of the nut thread when the jofnt slips under transverse force. Thus, the internal off-zorque force becomes sufficient to turn the bolt or nut completely loose as soon as the file tion is elfminated in the bearing area as well as in the thread tea. Such transverse slippages are more common in prac ice than is usually accepted. Experience shows that these joints most frequently fail by self-loosening. ‘An example of self-loosening is the fastening of the crown wheel m the body of a differential gear in an automobile. ‘The driveshaft teansmits the moment to the crown wheel by a bevel wheel and from there by grip friction to the body‘SELE-LOOSENING OF FASTENERS of the differential gear, Quite often the transverse slippage bepveen the crown wheel and body oceurs around the area Of these two teeth, which are actually in contact, because the adjacent bolt coanection is more heavily loaded than the more distant ones. Finally, in the Limits of the elastic deformation on the crown wheel, « transverse slippage takes place and forces the bolt to move as a pendulum in the tapped hole of the body. During the rotation of the crown whéel, one bolt after the other enters the ares of the ovo teeta in contacts thus one bolt after the other rotates loose. Also, in cases where the joint is axially or coaxislly loaded, tansverse slippage can occur because such loadings can cause elastic. defor~ ‘mation of the clamped parts (for example, a bearing cover). Proof for the validity of the whole theory of the loosen= {ng mechanfsm of wansversally loaded joints can be easily Getermined. SchoeUhammer, of the Laboratories of Daimler Benz AG., Stuttgart, hat determined the external off-rorque of several fastoners, including self-locking types under dy namic transverse loading. In these tests the frequency of ‘wansverse impact wat increased to the poiat that the exter~ nal off-vorque decreased, and the fastener finally rotated toose without the necessity of any external off-torque. For these tests he used a vibratfon machine, with the fastener to be tested clamping a plate against a block. ‘Two ait actuated impact hammers induced the relative motion between the plate and the block, and by variation of the als presse the induced impulses could be varied. ‘Tho bolts tested were M10 Grade 8.8 (equal to SAE Grade 5) with and without self-locking elements and also with pre~ vailing torque type self-locking nuts. The belts and nuts were tightened with @ torque wrench, but the actual preload ‘wasnot determined. Whena certain impulse with increasing frequency was reached, the external off-torque for loosening LSE SLR AGREE Oe oo short screw without Leck fom=-re short screw with spr ‘wather eee Leosening Torque [mst] los | i Per Fig. 5 - Loosening torque for preloaded fasteners under trans= verse vibration an became zero. ‘The nuts under test rotated completely loose by themselves. ‘A similar priniciple'was used in the earlier work of this author. It consisted of two plates clamped by the fastener to be tested, which in turn was transversely loaded dynam~ {cally by using a 6 ton fatigue testing machine. Transverse loads were measured by a tensile load cell, atid the ampli- tude of the relative movement between the two clamped plates (caused by the transverse loading) was recorded by a displacement pickup. Both sets of data were plotted by a high speed recorder. When tightened, both bolt torque and proload were measured by load cells. Thus, the bolt could be tightened to the exact preload. “The properly tightened joint was vibrated and then, dur~ {ng the vibration process, the bolt was loosened and the necessary extemal off-torque was measured. With {ncreas~ ing induced transverse force, the external off-torque de creased and, as Fig. 5 shows, the curve passed through Topp = 0 fora certain amount of induced vibration. Fuc~ ther increase of induced vibration resulted in negative teque values, indicating clearly that there is an internal off-torque doveloped by the wedge effect of the axially loaded thread. ‘The proof for the elimination of self-locking was given when fan internal off-torque of ~0.58 mkgf (60.4 in.-Ib) was re~ corded. This measured internal off-torque was approxi~ mately equal to the one obtained by 2 calculation for the case of a bolt completely free of friction: 8s Te Ey EA © ‘The same proof could be obtained by using a vibration de~ vice in whch a joint is clamped by one single boit in the center of a torsional vibration. ‘TO PREVENT SELE LOOSENING BY MEANS OF DESIGN ‘According to previous research work, {thas to be as~ sumed that total self-loosening does not occur in joints that ‘are dynamically loaded only in an axial direction, ‘This may be at east valid in all cases where the Joint is properly tightened. But ft hat been proven that at least partial rota~ tion of the nut can occur when the ratio between working load and preload fs high, and particularly ifthe load in the thread islargely diminished by the compression amplitude of an altemating working losd. ‘The partial loosening by rotation leads to a further oss of preload, and thus to & stop of the rotatlig process, because {he internal off-torque fs proportional to the preload. which Iwelf is reduced by the rotation process. On the other hand, the loss of the preload increases the danger of fatigue faflure because the total amount of alternating working load (axta1) is felt by the bolt. ‘On the other hand, total rotational Soosening can occur tn transversely loaded connections as soon 2 the clamping338 oad can no longer maintain sufficient grip ftiction so that transverse slippage between the clamped parts ocours. Trans= verse slippage, as a matter of fact, 1s quite common because ft is caused not only by transverse loads or transverse com ponents of a load in any direction, but also by coaxial loads that deform the clamped part elastically (that is by all joints that are loaded eccentrically, such as bearing caps) ‘The proposed concept of the mechanism of eelf-lootening cor seif-rotating of fasteners completes the picture of bolt failures under vibration by introducing the “self-rorating process.” Many cases of fatigue failures under obviously very low preload, which could not be explained by plastic deformations or even by insufficient tightening, were prob= ably caused by partial self-rotating. In Fig, 6anattempt ismadeto illustrate the failure mech- anism of dynamically, loaded bolted connections (7). This ‘chart has been made wp to resemble a road map with road numbers. There is a total of 12 different lines, which lead either to fatigue failure of the bolt (infrequently the clamped part or the nut) or to self-loosening of the fastener. ‘This ‘lustration shows that the basic rule for avotding bolt failure, either by fatigue breakage or by self-loosening. is t0 supply fwsurrcent Dine ‘ec insufficient vite ls ere revs ‘fe eattening 2 é TBEDDING _(SETZEN : | iy er ep sete LOOSENING, FATIGUE FAILURE ‘| Fig. 6 ~ Graphical filustration of failure mechanism of dy~ namically loaded bolted joints GERHARD H, JUNKER, and to maintain sufficfent clamping load. Since this paper deals with the "self-loosening” branchof the failure mech~ anism, the design mule for avoiding self-loosening or loss of preload by partial self-rotating is as follows: Design in such a way that no relative movement oceurs between the ‘contact beating surfaces normal to the axis of the bolt or betweon the tread flanls of the fastener components. ‘This requires that 2 sufficient residual clamping force remains whenthe scatter of preload values after tightening ‘and the embedding phenomena (Setzen), due to plastic de~ formations of the fastener or the clamped parts, is taken {nto consideration. ‘The residual clamping load should never fall short of the value F.,/# (wansverse load /iriotion coeffi= ‘cfent) to make sue that no tansverse slippage cau occur. ‘Modern rules for calculation of bolted connections (8 and 9) are therefore based on the necessary residual clamping force, the estimated degree of embedding (Setzen), and the scatter of the preload obtained by tightening, as represented in the following equation: Fax 7 AUg t By + OF ae yl @ where 1) ‘Tabulated values for the different factors are given in Ref. 8. ‘As itwas previously shown, using the example ofthe crown, wheel, there exist cases where these demands can be met only ff the joint fasteners are uneconomically overditnen~ sioned. There is no doubt that many dynamically loaded bolted connections are beyond an exact calculation; there~ fore the relative movements cannot always be excluded by ‘means of the joint design only. In these cases, wxo is re~ ‘commended of fasteners with self-locking etemeats that tend to resist the internal off-torque which occurs when inherent self-locking is lost. PROPOSED NEW TEST METHODS AND MACHINERY Since relative movements between clamped parts and fasteners cannot be avoided by means of the joint design only, and self-locking fasteners have to be used for pre~ venting “rotating loose," 2 criterion has to be found for the effectiveness of such self-locking fasteners. Several test method: have been tried in the past and are still in wse (LO 19). Most of them are more or less qualitative methods, ‘which result fa “rotating loose” or “not rotating loose" under certain specified conditions (some of them In preloaded condition; some in an unpreloaded condition). It {s obvious that the latter ones are practical only for prevailing torque types of locking fasteners. Other specifications give values Qa ‘SELE-LOOSENING OF FASTENERS for installation torque, breakaway torque, and prevailing torque. ‘With this paper an entirely different approach is proposed. Since tansversly loaded joints tend more to self-looseaing, the test procedure suggested imitates these actual conditions. ‘The first attempts were made with a device consisting of ‘two parts clamped together by the specfmea, with load cells and a displacement pickup 1 record transverse 1oad, preload, and displacement. The product, tansverse force times displacement, was ccalled vibration energy. Its maximum values were sigaifi- cantly different for the various locking elements. ‘This test method yielded results that could aot be te~ produced, since the machine used for the tests was a resonant type of fatigue machine. This type generates a force that fs @ function of the machine frequency. When starting a text, the transverse force increased with increasing frequency 23 a function of time until the loosening process of the tested fastener was initiated. Depending upon the starting speed of the machine, a different maximum force (and therefore maximum energy) was necessary for rotatton loosening of ‘a autor bolt. Therefore, anew vibration machine was de~ signed for these tests. ‘The prototype of the new machfne was bullt by the Cor~ porate Machine Building Dept. of the SPS Co. This ma~ chine generates a mansvesse sliding motion between two ‘clamped parts by means of an adjustable eccentric. The resulting transverse force is independent of frequency and starting speed of the machine. It is for a given machine stiffness (pring constant of the load twansmitting part), suletly @ function of the eccentric adjustment. Figs. 7 and 8 show the heart of the machine. "The tested bolt clamps the U-shaped top part onto the bed, an integral omen CONNECTING PLATE To LOAD CELL FLEXIBLE FOR TRANSVERSE SHAFT FORCE SPECIMEN DISPLACEMENT plek=Ueqvon) ~ Sy COMPRESSION LOAD Fig. 7 = Central part of the vibratfon machine sig part of the machine feame. ‘The clamping force is meatured ‘with a compression load cell through which 2 bushing with {ncemal threads is placed for testing bolts. When testing nuts, this bushing may hold e threeded std, or a special ‘bushing with external threads may be used. ‘The U-thaped top partis displaced parallel to the bed. ‘The force is transmitted by a load cell mounted on flexure plates, 2 connecting rod being attached to the load cell by crosswise flexure plates and by an adjustable eccentric. The ‘U-shaped top partis soparated from the bed by-flat stipe of needle bearings, to avoid galling. ‘The relative movement between bed and top part is mea= sured with special Linear differentisl transformers. The ‘ughtening and loosening angles were measured with a Mnear potentiometer, which was attached to the test specimen by a floxible shaft. Exchangeable threaded bushings and inserts in the U-shaped top part make the mechanism ussble for testing bolts from 1/4 in. (or M6) up to 6/8 in. (ar M16) dead sizes. "This range covers the majosity of the common locking elements. ‘Tho machine is driven by a 9.6 KW (6 hp) a-e motor, whfoh ts located in the machine housing. & V-belt drive and dffferent pulleys allow testing frequencies of 1500, 3000, 4500 opm. Pig. 9 shows three different applications for the machine. In the upper picture the U-shaped top part is connected 10 the load cell with a bolted plate, to eliminate all play. ‘The arrangement is used for tests with pure transverse Tosd~ ing. This plate is replaced in the middle picture by an U-shaped theet metal specimen that transmits the transverse forces when clamped with the U-shaped top part tothe bed. With thi acrangement the fatigue sirength of bolied sheet metal joins ean be evaluated (See also Fig. 21). This ts of special interest if the fastoner damages the surface of the clamped pars, for example, when usiag locking fas~ reners with serrated bearing surface. In che bottom picture the connecting plate is replaced by s connecting rod. It Is thereby possible to rotate the bed so that it forms any Fig. 8 ~ Photograph of central part of the vibration machine320 angle between 90 and 180 deg with the frame of the ma~ chine, The center of rotation coincides with one bearing of the second conneeting red. ‘This arrangement allows load~ ing the test specimens by any combination of twansverse and axial forces. ‘The recording equipment used for the tests described above is shown in Fig. 10. Transverse force F, and dis- placement “d" are plotted on a Visicorder as a function of time. ‘The bolt preload is recorded on a third channel. By using a two channel oscilloscope, the tansverte force F, fs recorded as a fnetion of displacement “a” in X-¥ hookup ‘on one channel and the preload on the second channel, Run~ ning the tests with a special low speed drive (10 cpm), the same information was plotted on an X-Y,-¥, recorder. ‘The electrical values of transverse force F, and displace ment “d" were multiplied by a specially built electronic rmultiplication cireuit. The result Fd fs plotted as a func ton of preload on an X-¥,-¥, recorder. The tightening ‘and loosening angle measured with a lnear potentiometer GERHARD H, JUNKER 4s recorded as a function of preload on the second channel of the X-¥,-¥, recorder. “fo check out the machine and to evaluate the posstbili= sfes and limitations of the various testing, measuring, and recording methods, the test specimens were restricted to nonlocking screws of different thread tolerance (with and srithout spring washers) and to tvo types of free spinning self-locking screws with seated bearing surfaces. This restriction has been made because of lack of time and be~ ‘cause these types of screws are used in great numbers for sheet metal constructions, particularly in the automotive fndustry. The described test methods and machinery can ‘be used for testing all kinds of self-locking nuts and also self-locking screws of the provatling torque and adhesive A comprehensive study with specimens of all groups of self-locking fasteners is in preparation. Geometric shapes i =e ¥ L ‘WSICORDER RECORDER See i (* Fig. 9 - Three different possfbiliies of application of the vibration machine oao ceut cone, aD ISPLACEMENT Foerrmatevense CELL FOR Pree uP (oT) Fore Fr PReLonD Fy FORE Fig. 10 - Plan of instrumentation and recording & e‘SELF-LOOSENING OF FASTENERS and dimensions of the test specimens ate'given in Table 1 and the mechanical properties in Table 2. Screws of the size M10 X 1.5 and 8/8-16 have almost equal nominal thread diameter, pitch diemeter, and pitch so that their test results can be compared. All fasteners were oiled before testing and were used with plain washers under the head, having 2 Rociwell hard= ness of TTR, (By, * 187 gf/mm?). ‘The free spinning © 324 types of locking fasteners were additionally tested on washers of higher hardness (R, = 99 - 41). During all tests the specimens were criginally preloaded to 2500 lgf = 5500 Ib. This relates to 5% of the proof load of sorows with a strength level of 150 Class 8.8 (similar to Grade 5). All tests except those run with a special Iow speed drive (10 epm) were run with a frequency of 3000 cpm. X+Y PLOT OF TRANSVERSE FORCE AND DISPLACEMENT ‘Table 1 = Dimensfonal Spectfication of ‘Tested Specimens Specimen A. Nomlocking Screws ‘A, Hex head cap screws, DIN9S3, MIO x 1.5 90 Socket head eap screws, 1960 Series, 8/8 “16 X 1-1/4 in. 2 Socket head cap serews, 1960 Series, 8/8 “16 X 1-1/4 in, B Spring Washers DB, Hex head cap screws, DIN 983 (4,) plus spring washer DIN 128 C Free Spinning Locking Screws Cy, Flange head. hex cap screws, ML0X 1.5 X 80, with iangular teeth, ggoove and bearing plat form near the shank C, Flange head hex cap screws, MOX 1.5 X 30, with Long radial teeth and circumferential bearing ring, edges of teeth and bearing ring forming one continuous curve O Dimensional Specification allowance: 0.944 mm. 0.01855 in: Pitch 1.586 mm ‘Thread tol: Bolt 24 Nut 28 Nominal max. theead allowance: 0,290 mim. 0.01142 tn. Pitoh: 1.586 mm ‘Thread tol: Bolt 3A Nut 33 Nomfnal max. thread allowance: 0.199 mm 0.00759 in. ‘Thread tol: Bolt 6g, Nut 6H Nominal max. thread allowances 0.344 mm 0.01856 in. ‘Thread tol: Bolt 6g Nut 6 Nominal max. thread * allowance: 0.944 mm 0.01865 in.302 (GAYSTERESIS CURVES) - The sghematic drawing in Fig. 11 ‘explains the relationsh{p of force wo deflection and to dis~ placement in a slipping, bolted’connection subjected 1 dy~ namic transverse forces. The dotted line represents the force /displacement diagram of two plates, which are clamped together by a screw with narrow thread tolerance. Iwill be assumed that the head of the screw cannot slip of the plate. ‘This serew is bent to the left by the mop plate at point A, eft dead cente:}. When the top plate moves to the right, the setew bends to the right until (owing vo the moment in the thread) the slipping in the thread begins at point B, + the force /deflection line is less steep up to polat D, . AED, the allowance in the threed has been used up by slipping and the screw thread is supported by the opposite side of the intemal thread. From there on to point B, (ight ead center) the screw is bent more. ‘The same process re peats during the other half of the eyele when the screw is bent back to the left dead center. The included area rep- resents the friction energy absorbed by the joint, which 1s being transformed into heat. The srea is very small, since only the thread slippage creates friction energy: the bending of the sorew is pure elastic deflection. ‘The curve, indicated by a dash-dot line, represents the force /aisplacement dlagram of a bolted joint for whfel the bolt head toocannot slipon its bearingsurface. The thread tolerance, however, isso wide that the external serew thread (GERHARD H, JUNKER is contained by the internal thread exactly at the left and right dead center. This means that the slipping in the thread starts at point B, and ends at point ~The enclosed area 4s larger, owing to the longer slippage displacement in the threads; therefore the absorbed fction energy is latger £20 “The soli line in the dingtem represents a joint with a fastener that can slip under the head. ‘This is the case when the flction grip Fy! ie overcome by the wansvere force ge Starting from the left dead center a point Ay the sexew is bent to the right by the transverse force, witch does not yet break through the friction grip of the head and tne top plate. Ar point 8, slipping in the threads starts. Before thred slipping can proceed to D:, (for narrow thread roterance) or ', (for wide tread tolerances), sipping under the sorew head stars at point Cand ends at polat fy the right dead center. Drawing aes trough D', and D', parallel wo the oad/ deflection line of the load wansmitting machine part (C,, = tan @), we obtain the distances W, and W., a6 te Intersce= tons of these parallel Lies with kino C,E,. They represent the part where slipping in the thread as well as under the hhead occurs. Fastener rotation can take placeonly on zhese ‘stances because only there is the sorew free of friction. ‘Table 2 ~ Mechanical Properties of Specimens BV, 3 mroprtonsl . Pm ggg, amon S00 ne _tgtfrom?—saffom site Nor __pal pal Gore Hardnese Case Hardness Remarks a 0 pe.coo 146.600 . ow oso seq 166.400 he oan so reac 166400 ate 2 a0 ces Ml, popes of 9800 116,800 aes cm % os 0s Cam hasdaam ne aias0o 188.700 ned (.08 mm (os00197 ta) rom ‘itace _ 8 oe 20 s20 soso 198400 ( $sELE-LOOSENING OF FASTENERS mont DEAD ceNTE; (2\ osseuscenen & LEFT DEAD CENTER. ‘woe THREAD FIT srresot wneans ny “Ao Srrtas (MOEN HEAD CY si srr sr es ‘wh fom close TREAD Fit wi Fon woe THREAD Fi Fig. 11 ~ Schematic description of the self-loosening process under tansverse vibration s28 It can be soon that the distance W, for threads with close tolerance is smaller than the distance W, for treads with wide wlerance. ‘Therefore, screws with close thread fit ‘must have a higher resistance. to self-rotating loosening than screws with wide thread fit. These considerations ex= plain that the prevailing torque type of locking screw works by reducing or eliminating the thread allowance by various ‘means. The self-forming screws must be added to this cat~ gory for the seme reason, ‘According to Fig. 11, a bolted connection that slips under the bolt head will absorb the most energy: the hysteresis loop encloses a considerable larger area. Where the clamped parts are not separated by flat strips of needle bearings and ‘where they are clamped together ditectly, the interaction of force, deflection, and displacement is basically the same. Merely the fefction energy besween the clamped pare (Fy 2d.) has to be generated additionally every cycle. When the transverse force F,, is large enough t force a relative Aisplacement of the clamped plates, the same process as that described in Fig. 11 takes place. ‘The use of strips of needle bearings between the clamped plates merely prevents galling: ft does not favor either group of locking elements. ‘The diagrams recorded during tests have basically the same shape as the schematic drawing of Fig. 11. Fig. 12 shows the mansverse force-displacement diagram of a non- locking screw A, and of a free spinning iocking screw Cp, recorded at a machine speed of 10 epm. The diagram is presented for 100, 80, 60, 20, and 0% of the original bolt preload, which is also shown on the same plot as a function of displacement (upper part of chart). Tt can be seen that SSS forefoot SP i: a —1 foeeetered : ease s om Serene feo semee yf ee bubua a uabn buen ait : ; iat 300 te S At « ff Se aie T ® 2005 ment "a" (hyteresls curve) for 500 1 400- acer See a easueal ics924 the nonlocking serew A, starts t slip under the head in the fully pteloaded condition; however, the screw C, with sex~ rated head bearing surface does not start to elfp until the preload is decreased to 40 of the original preload. ‘The tangent citves in Fig. 12 represent the load /deflectfon curve (Goring constant C,,) ofthe Toad transmitting parts of the machine. Fig, 18 shows the complete loosening process of a non~ locking screw A, recorded with an X-¥ recorder at a test frequency of 10 epm. ‘The recorder was turned off at a re~ maining preload of 10% of original preload, but the screw had completely turned loose after 200 cyéles. Fig, 14 shows the same process for the same eccentric ‘adjustinent (zero load amplitude d, = 20.8 mm), but re ‘corded on an oteloscope during the 9000 epm test fre~ (quency. Both diagrams, recorded at10 epm and at 3000 cpm, are equal whon the different scales are taken {nto considera~ tion; even the number of cycles causing complete loosening {s approximately 200 for both speeds. This result. gained 1m several tess shows that the self-loosening of screws is independent of feequency. 1c simply depends upon the oo~ ‘currence of relative motions and on the length of such mo- tons during one cycle. 1f the forces causing thete relative rotions are inertia forces, they are a fonction of the square of the frequency. In this case the frequency indirectly in~ fluences ihe Toosening process. ‘This test method yielded interesting results when record~ “SPECIMEN At fig, 10 ~Proload F and tansverse force F versus displaces Fig, 14 ~ Transverse force Fy versus displacement ment "d", showing a continuous self-loosening process of a nonlocking screw A, ; frequency 10 cpm GERHARD H, JUNKER {ng the angle of rotation during loosening fa aditfon to the force /lsplacoment hytaeds curve. Itbecomes apparent () hae the looting proves of « nonlockingfartener stares with the very fits eycle. The recorded angle of mistion Tootening was 0 deg after 25 eyeles: i was constant, since 86 deg rotation occurred per eycle. The recording of preload and angle of rotation asa function of displacement duving 10 cpm tess gave an answer to the question of what causes the relative high los of preload of el fe spinning, self-locking fasteners atthe beginning of te vibration pro tou. This Phenomenon was especially encountered attest series descibed late. Fig. 15 shows the first 60 cycles at 20 cpm frequency for free spinning locking reews C, and G, (te fe five cycles are recorded, then every fifth; sorews seated on a soft washer). The screws lose 25% and 14.4%» respectively _ of thefr original preload during the test. Recording the angle * of rotation showed that the high loss of preload at the be- ginting was not only the result of the “digging in" of the teeth but also of a rotation of 1.8 deg for C, and 0.6 deg for G,. After 80 cycles, the rotating process stops and ur~ ther les of preload is due to briaeing ofthe teeth Sato dhe sutfaes. “The loss of preload aftr 50 cycles Is the samme when the sere are seated on hardened washers; however, it is caused toa lager dogaee by rotation. ‘That ths proces is dependent largely on the shape of the teeth becomes obvious by the following: The serewe C, having long rectangular tect ost. show- ing a continuous self-loosening process of a nonlocking screw ‘A; frequency 8000 epmSILE-LOOSENING OF FASTENERS only hatf as much preload after 60 cycles under equal con= Gitions and tuned only half the angle as the screw C, » which hhas triangular high teeth. Fig. 15 (C,) aio hows clearly dhe two bends in the ys teresis curve which characterize the limits of thread slip~ page. The hysteresis curve recorded by the X-Y recorder hhas the same shape as the one drawn in Fig. 11, which shows fa screw that does not slip under the head. This kind of test {s primarily suited for evaluating rotating and brineling pro ‘cesses, Tt can be used expecially well at slow testing speeds ‘so that X-¥ recorders may be employed. It is not so well suited for large test series at high testing frequencies when ‘statistical evaluation is desired, siuce in that case the hysteresis curves will have to be photographed from an os- eilloscope. Fig. 16 shows meximurn output of the vibration machine and actual hysteresis curves. VIBRATION PRODUCT ~ A fastener connection that slips under alternating tansverse force absorbs exactly that “amount of energy represented by the included area of the hysteresis diagram. Assuming the extreme case of an 2b~ solutely stiff screw. all absorbed energy would be in form of frietion and would be for 1/4 cycles By 7 aFp @ [At the other extreme, pure beading of the screw (n0 feiction), 1/4 eyele, would absorb © tn the sgeond case the energy for one full cycle would be ‘ero. ‘The largest possible energy per 1/4 cycle for a given oe og 04 02 | 02 04 4508 2500'% 05 Fy (kof) SPECIMEN C1 325 eccentric adjustment would be the following rectangle, ex~ wren by 1 ty (2) -e a) \e) 9% “The Limiting values of displacement and transverse force, the execeding of which causes the preloaded screws to rotate, toose, will be used for the rating of locking abilities of var~ fous locking elements. Both values ee of equal importance, since & slipping joint will be kept from total separation tf targe displacements in connection with small forces, es well as large transverse forces in connection with relatively small ‘isplacements, are absorbed. It is therefore propoted to ine troduce the product of transverse foree times displacement 6 ata erlterion for te locking abity of a osking ele- co) ment subjected to tansverse displacements. will be called “vibration product” V. ‘The vibration product V was obtained in two ways: First, displacement “a” and transverse force F., were recorded on ‘The product fa Visicorder together with the bolt preload. Values for “a” and F, were then read from the Visicotder chart and malti= plied. As # second method, the product was obtained by electrically multiplying the signals of the two amplifiers of displacement and mansverse force, The multiplication ‘citcult was built from commercially available components, By muliplying "é” and E,eleotteally tt was possible ‘tp record the vibration product as a function of the (de- ‘ereasing) preload on aa X-¥ recorder, Simultaneously, the ‘ightening and rotation loosening angle was registered as ‘a function of preload on the second channel of the X-¥ re~ 02 Of O8 08 Fig. 15 - Preload F,, and trans~ ‘vere force F., versus displace ment "2", showing the first 60 ‘eycies of vibratfon for free spit ring self-locking setews C, and SPECIMEN C2 —_C,; frequency 10 epm526 9) Caton ae ff Sy Ears # ote Cae 918 Kgt/men ®) oat, 3 ar Cn \ Fre *Frn dn eB Ears 92 ty = 2Ey Enett™ Ent Gp dy?= 52. ay? dp VE-4y Feat VE Fre Fig. 16 ~ Maximum output of the vibration machine and actual hysteresis curves (GERHARD H, JUNKER cordet. Recording the angle was done in the following way First the serew was tightened to 500 kgf (1100 1b) preload ‘snug tight); then the screw head was connected with the angle recording potentiometer and the screw subsequently tightened to 100% preload. Fig. 11 shows recordings obtained, as described above, of the vibration product ané the tightening and rotation loosening angle as functions of the preload. ‘The machine settings were a8 follows: eccentric adjustment to produce Ejay 7 210 mm gf, frequency = 9000 cpm, no. of eycles = 8000 (provided total loosening did not ocour). Fig. 117 shows aio the performance of screw C, after one tightening, tested ‘at 3000 opm frequency and an energy level of E,,,, = 210 mm kgf. When seated on 2 soft washer, the screw retains ‘50 of its preload after 2000 cycles, whereas the vibration product hardly changes. The anglo A ~ 6 dog, ropresonting ‘the rotation loosening, and the angle B corresponds to the ‘amount of embedding during vibration, ‘These can be de~ ‘termined when the line of the tightening angle of a plain unserrated screw of the same dimensions fs added to the recording in Fig. 11. If this is éone, one sees that the sorew oses 1% preload by partial rotation loosening and 80% by ‘embedding, 2 total of 0.08 mim (0.008 fn.), which corre sponds to the angle B. The total angle of tightentng is 84 deg; this is S¢ deg (C) more than the tightening angle 50 deg (© of an unserrated screw of eaual dimensions. The differ ence of 34 deg represents the embedding and the elastic deformation of the flange during tightening (9.15 mm 0.006 in.). The same picture shows the loosening process of the seme screw C, under the same. conditions except that it is seated on a hard washer. The screw rotates ap- proximately 60 deg and loses all its preload. inally, Fig. 17 glves the pesformanes of a serew C,, Which is of the same basic type. Loss of preload and vi~ bration product are the same when seated on a soft washer; however, the tightening angle is only 40 deg compared with 2A deg for G. ; the reason being a less deep embedding and 2 smaller elastic deformation of the flange. ‘The rotation ‘Toosening angle of 6 deg is approximately the same. A sig- nificant difference shows up when the screw is seated on hard washers. In this case the serew C, rotates loose only 6 dog after 9000 cycles, the shape ofthe angle dlagram {ndlcating thatthe loosening proces has come m est after eg. Many more tets were run inthe sare manner with the fasteners reported in Table 1. They were evaluated at dif- ferent energy Levelt (cero load armplinide adjustments or eccentric persion), recording preload, angle of rotation, and vibration product, te latter belng the electronically Mieco tuciect einai est =) Only a few of them ae reported tn Fig, 17, to deserthe the test procedure. This mode of esting gives a quick basle nowledge of the peeformance ofa locking element, a duadvantage beg‘SELE-LOOSENING OF FASTENERS RELOAD VORSPWRERAFT Fe ig) Ctr moen EE cee |, ea al 4] \F “t =" i 5 a

produce the zero load amplitude d,, which yields the de~ : eee: ViBRATION=PREDUCT Fy = ¢ VIBRATION-PRODUCT Fy od PRELOAD Fy [Fro] UI Tae ¢ fll gti eH g | lil | ; Lt fat ti 4 d | IL i ‘ i 1 | ee ; 2 Number of Cycles Fig, 19 ~ Vibration product and preload versus number of cycles (VN and Fy Number of Cycles IN curves) for specimens C,, C,380 GRRHARD H, JUNKER oy) ” wo 1000 0 1 1000 eD Fig. 20 Preload curves versus the number of cycles for varlous maximum enesgy level E-F-N curves for specimens By Ay By By€ ‘SELF-LOOSENING OF FASTENERS seed energy level acoordag w the equation (4) @ a) % As presented in Fig. 16, results gained from different E ‘max ‘machines must be comparable, even if the spring constants of the load transmitting parts are in the ratio of 1:2. Fig. 16 (4) shows that transverse vibratfon tests, run at equal energy levels nae) * Ea max)’ 44408 Machines wth spring constants in the ratio of 1:2 resulted {n comparable force / deflection (displacement processes. If, as reference for the text levels, the transverse force or the zero load amplitude (eccentric adjustment) were used in the Case of two ma~ cchines with 1:2 spring constants, the results would be very different force deflection displacement processes; see Figs, 16(0) and (6). a2) — 331 ‘The procedure to generate E-F-N curves is as follows: 1, Adjust the zero load amplitude with the eccentric of the vibration machine to obtain the first desired energy level: F max M 2, Proload the specimen to 15% of proof loads read pre~ oad off bridge amplifier. 3 During vibration process, read numbers of eycles cor responding to 80, 50, and 20% of preload F,, as indicated on bridge amplifier. 4, Repeat process at lower energy level until the preload does not drop below 80% after @ predetermined number of cycles. Run several tests at each energy level, similar to fatigue testing, so that a statistical evaluation of the results 4s possible. as) C1 soft washer C1 hard washer , (emia) C2 soft washer Fer Tanke) C2 hard washer Fig. 21 ~ Preload curves versus the number of cycles for various maximum energy level E-f-N curves for specimens ,, C, om hard and soft washers392 SUREACE INTEGRITY - The value of self-locking ele~ ments with serrated head bearing surface is often diminished by the damage they cause to the surface of the clamped parts, despite their excellent locking performance. They ‘itt nothes in the surface of the clamped parts, which reduce the fatigue strength of sheet metal designs as well as of other ‘machine components. ‘The new vibration machine can be adapted to evaluate this influence, t00, as is shown inFig. 9 Fig. 22 shows the teat seuup for 2 shicet metal spectmen. "The specimens used for the tests desorbed here were ‘made from 1 mm (0.04 in.) cold rolled body sheet, bent nto a U-shaped channel. The sheet metal specimens were ‘ested with two types of screws with serrated beating surface (G, and C.). For comparion, txt with nonlocking serews ‘were run. ‘The fasteners were tightened twice to 100% pre~ i z T | | r g rei eae umber of cyelan Fig. 29 - Transverse force versus number of cycles, F..-N curves for clamped sheet metal specimens GERHARD H, JUNKER: Joad, unscrewed, cleaned, and seated for a third time to 1106 preload prior to vibration testing. After taking out the flat stipe of needle bearings, the tests could be run like standard pushpull fatigue tests. ‘The results are shown in Fig. 28. Some typical fatigue failures caused by sorew C, are shoven in Fig. 24, and those caused by screw C, in Fig. 25. These tess make clear the difference in the shape of the serration of the two screws. , cus an annular notch, angular in cros sections, a5 ‘can be seen in the washer left over from other tests (Fig, 24). ‘The fatigue failure ocours along that snnular notch; as = result, a washerlilee piece {s cut out by the serew. The screw has long radial teeth, the edges of which run out con~ tinuously into a smooth outer bearing surface. The damage done to the seating surface is minute, and therefore fatigue failures oceur away from the fastened area, as is shown in Fig. 25. For this reason, the F.,-N curve of serew C,, is close to the S-N curve of the undamaged sheet specimen. Fig. 24 - Fatigue fatiure of sheet metal specimens clamped ‘with free spinning self-locking serews C, 7 shape of teeth and removed washer after vibration Fig. 25 ~ Fatigue failure of sheet metal specimens clamped ‘with free spinning self-locking screws C+ shape of teeth and removed washer after vibration 9‘SELE-LOOSENING OF FASTENERS CONCLUSION AND FUTURE ‘TEST PROGRAM Self-loosening characteristic of fasteners can be achieved with a new vibration machine that generates relative motion in the clamped parts perpendicular to the axis of the fastener. Several values characteristic of the process can be measured during the test. These {nclude preload, transverse force, displacement, and angle of rotation, The described ma~ chine is suited for further research wotk to study the mech ism of self-loosening. It must therefore be equipped with a series of electronfe recording instruments. “The machine is also suited for large scale testing and inspection testing. For this kind of work, only one bridge amplifier {s needed to record the preload. Whea so equipped, E-FsN curves can be recorded for large scale and inspection tests, Thereby a quantitative testing of locking performance of all Kinds of locking elements becomes possible. The fasteners are tested in 8 preloaded condition. Test results of different locking types of fasteners such as thove having prevailing torque, and free splaning torque, and adhesive type screws can be compared. ‘The spring constant of the machine must be known. Results obtained from machines ‘with different spring constants must be comparable. ‘The following details must be resolved preparatory 10 the specification of a standard test method: 1, Influence of serew length, 2, Performance of nuts tested on special externally threaded thserts or suuds. 3° Vibration endurance Umit; that fs, after how many cycles has the self-rotation of a fastener or the loss of pre= oad eojite to rest. 4, Tnfence of hardness of clamped parts; specification, of standard test washers. 5. Influence of thread tolerance. 6. Influence of thread pitch. 1. Reusabilit Of spectal interest fs the determination of dhat number of cycles after which no more self-loosening occurs. ‘The ‘exact shape, particularly the horizontal part, of all F,-N, oN. -N, and F.,-N curves presented in this paper is not yot known because only a Iimfted number of tests have Deen run so far. Tecan only be estimated that the "vibration endurance limit” is smaller than the fatigue stength of ma~ terials and structures. Furthermore, tests should be run to establish the influence of superimposed forces other than transverse forces. First screening tests using angles of 15, 30, 45, and 60 degshowed that under an equal external force, serews loosened faster when tested under smaller angles. Itean be assumed that the pure transverse force applied at an angle of 180 deg represents the most severe conditions. The influence of Inbrication, tolerances, and surface finish should be investi~ gated. The question is whether there is a correlation bo ‘reen static tests (ixcaktaway torque and s0 on) and vibration 333 tests. ‘The answer to this will probably be found by statistical methods. When enough tests have been run to make it possible to give statistically evaluated data concerning the locking performance of locking elements (@-F-N curves, endurance limits for F, = 86, 50M, 20%), then a truc analysis of the value of locking elements will be possible. In addition to the then measurable value of locking performance the following factors must be taken int consideration; reusability (a), surface integrity and fatigue strength of clamped parts (6), fatigue strength of clamping parts (c), compatibility with clamped material and hardness (é), installation cost (e), ‘and price (f). Factors (a) and (6) can also be evaluated with the desertbed vibration machine and the proposed meth~ ods. When a catalogue with E-P-N curves is established by 4 large number of tests, short term tests with specific energy Jovels canbestandardized as inspection tests, the minimum number of cycles for F,, = 80%, SOF, or 20 have to be specified. It was not intended to give at this stage of study an ‘evaluation of the locking performance of the tested speci- ‘mens. The results, therefore, shall only be considered as 2 ‘means for the demonstration of the operation of the vibra~ ‘tion machine and for the discussion of several possible test methods. A value analysis covering sIl the above mentioned factors will show which locking element is the optimum one for the actual application. SUMMARY Proloaded screws (or nuts) rotate loose, a5 soon af a rela~ tive motion in the thread takes place. This motfon cancels ‘the frfetion grip and originates an inner off-torque propor~ ‘ional to the thread pitch and to the preload. The inner off= ‘torque rotates the screw loose if the friction under nut or bolt head bearing surface is canceled by relative motions. ‘A dynamicatly loaded joint fails in most cases either by fatigue or by rotation loosening of the fastener. Even the fatigue fatlure is often initisted by partial loosening. If tt {s impossible to avoid relative movements by design, self- locking screws must be used, This means either that relative ‘movements in threads must be diminished or avoided so that the formation of the inner off-torque will be thereby pre~ vented, or that the formation of inner off-torque must be blocked up in the head or nut bearing surface. Dynamic transverse forces are more dangerous than dy~ namic axial forces. Axial forces cause relative movements through expansion of the nut threads transverse forces cause relative movements through rocking action of the sorew in ‘the internal thread (or rockng motions of the nut on the extemal thread). The relative motions caused by transverse forces are larger. No existing dynamic test method measures the locking performance of locking elements. A new vibratfon machine fs desoribed which generates transverse forces and displace ments in preloaded joints as well as combinations of tans~334 ‘verse and axial forces. Displacement, transverse forces, protoad, and angle of rotation can be measured. Additfon~ ally, the fatigue swength of bolted sheet metal joints can be evaluated, which 1s a means of rating the damage that flocking element causes to the surface of the clamped parts. ‘Various test methods using different recording instruments axe described and discussed in relation to nonlocking and self-locking screws of sizes M10 and 8/8 in. NC. The in= fluence of thread fit on the loosening resistance of noalock- ing screws, the influence of the shape of the serration of free spinning serewe on locking performance, the fnfluence fof frequency on the self-loosening process, and fatigue strength of clamped sheet specimen were studied. Finally, a simplified method for large soale testing and inspection testing {s proposed for which only one bridge. amplifier is needed with the vibrationmachine. Thismethod produces energy/aumber of eycles curves (similar to S-N curves) with 20, 50, and 80h preloads as parameter @-F-N curves). This test method could be developed into a stan~ dard test method after solving some detail questions. Tt ‘would yield measurable values for the Locking performance of nuts and bolts under dynamic loading. It could become the basis of a real value analysis for the use of locking ele~ ments. ACKNOWLEDGMENT G. Meyer, Labs Manager of SPS-UNBRAKO's European RE Dept., has contributed to this paper by arranging the ‘electronic test setup, supervising the voluminous test series, ‘and assisting in preparing che English version of this papers which originally was written in German. ‘SYMBOLS AND ABBREVIATIONS Displacement of elamped components under twansverse force, mm or in. Displacement of U-shaped top part of vibra~ tion machine in unloaded condition (ze20 oad amplitude), mm ot in, a, = Major diameter of thread, mm or in. a, = Pitch diameter of thread, mm or In, 5 = Dilation of aut under axial load, mm z = Amount of embedding @rineling, "Setzen,” plastic deformations) during dynamic dy name loading of a joint, mm or in. a = Tightening or trque factor, which considers the preload deviations caused by differ~ ences of filetion coefficients and applied torque in the design equation for bolted connections 4 = Spring constant of load transmitting parts ‘of vibration machine, Fig. 16 (ai hef/ am or Ib/in. GERHARD HI, JUNKER = Spring constant of clamped plates or com= ponents, kgf/mm or 1b/in. = Spring constant of sorew, kgf/mm or tb /in. = Friction diameter of a bolt head or aut, mm or in. = Energy for elastic deflection of bolt in a transversely loaded joint during 1/4 eyele, mm kgf or fn.-Ib = Energy absorbed by friction in a sliding foint during 1/4 cyclé, mm kgf or in.-Ib = Maximum output of vibration machine for ‘acettaineccentric adjustment, mm kgf or in-lb sxial force in a fastener, kgf or Ib ‘Maximum axial force in a prestressed fas~ tener under working load, kgf or Tb Rg = Residusl clamping force necessary for main~ taining grip friction of clamped compo- nents, kgf oc Tb ‘Transverse force in a bolied joint as well as in the vibration machine, kgf ot 1b F = Transverse force in vibration machine with 0 toventric adjustment dy kgf orb Fy = Preload in fastener, kgf oc 1b Fy = Working oad in a bolted joint, tgfor 1b ary Los of preload in a bolted joint by em- bedalng, laf or 1b L = Weight, gf oF 1b ° 1 Tranavowe fore for moving loads resting on 2 horizontal plane and sope, ligt or Ib Toee = off-nrque for a tneaded fastenee under static conditions, m kf or in.-Ib v = vibration product equals displacement “8” times transverse foree Fn a wansversly vibrated joint, significant factor for the locking performance of «self-locking fastens, mm igf 0: in.-tb « = Angle of loed deflection curve of 1oad trans~ rltting machine pars, dog ’ = Slope angle and hetlx angle, deg # Goefficlentof inictionintnreads, = tan p My CGoefficfent of fiction in tsads, under ‘head or nut E*E-N curves = Curves of constant preload for certain vi~ bration energy levels versus number of cycles EgeN euves = Transverse force versus nusaber of oyeles significant for surface integrity of clamped ‘components g, jy-N ourves = Decreasing preload under transverse vibra~ tion versus number of cycles V-N curves = Vibration product versus number of cycles c 4 WZSELE-LOOSENING OF FASTENERS REFERENCES 1. J.N. Goodier and R. J. Sweeney, “Loosening by Vi~ bration of Threaded Fastenings,” Mechanical Engineering (December 1945), 798-802. 2. J. A. Sauer, D. G. Lemmon, and E. K. Lynn, "Bolts How to Prevent Their Loosening," Machine Design (August 1950), 188-199. 3. E.G. Paland, "Untersuchunged ther die Sicherungsei- sgenschaften von Schraubenverbindungen bei dynamischer Belastung” (Iavestigation of the Locking Features of Dy~ namfcally Loaded Bolted Connections), Dissertation TH, Hannover, 1986, 4, G, Junker and D. Strelow, "Untersuchungen iiber die ‘Mechanfk des selbstttigen Lésens und die aveckmBige Sicherung von Schraubenverbindungen” (Investigation of the Mechanism of Self-Loosening and Optimal Looking of Bolted Connections}, DRAHT-Welt, Fachbeilage Schrauben ‘Mattern Formteile, Teil Iz $2 (1966), Ne. 2, 108-114; Teil tz 52 (1966) Ne. 9, 175-182; Teil Il: 52 (1966) Nr. 6, 617-885. 5. G. Junker, "Betrachtungen tiber das selbstitige LSsen. und die zweckmiBige Sicherung von Schraubenverbindungen” (Aspects of Self-Loosening and Optimal Locking of Bolted Connections), Maschinenmarkt, 72 (2966), Nr. 76. 6. G. Junker, “Warum ldsen sich Schraubenverbindungen ~ wie Kann man ¢s verhindern? (Why Fasteners Come Loose = 336 and How to Prevent it), industrie-Anizetger, 61 (1867), 19-22. 1. G. Junker and G. Meyer, "Neuere Betrachnungen liber ie Haltbarkoit von dynamtsch belasteten Scheaubeaverdin~ dungen” (New Aspects on the Fatigue Behavior of Dynar~ {cally Loaded Bolted Joints), Draht-Welt, Fachbeilage Sch rauben Mattern Formtefle, 54 (1968), Nr. 7, 487-499. 8. G. Junker and D. Blume. "Neue Wege einer system= atischen Schraubenberechnung” (Moder Rules for Calcula~ ton of Bolzed Connections), Disseldorft Triltch, 1965. 9. G. Junker, “Neue Prinzipfen der Schraubenberechnung, (New Principles of Caleulation for Bolted Connections), Mas~ hinenmarkt, 71 (1965), Nr. 74, 16-29. 10, AN-N-L0a, "Nuts; self-locking, 650° E." 11, AN-N-5b, "Nuts; self-locking, 250° 12, MIL-N-167817, "Nut self-locking, 1200°F." 12. MS 26531 (ASG). "Vibration Test RIG." 14. NAS 9356, "Nuts: self-locking 450°F, High Quality." 15. MIL-N-25027C, “Nut; self-locking, 250°F, 450°F, and 800°, 125 KSLFTU, 60 KSIFTU, and 30 KSEFTU," 16. MIL-F+18240, “Self-locking element for faetener ‘externally threaded 250° F." 17. TET-100, "Prevailing-Torque Type Steel Hexagon Lookauts." 16. IEI-101, "Torque~Tension Requirements for Pro- vailing Torque Type Steel Hexagon Lockauts.” 19, MUL-STD-1812, "Fasteners, Test Methods.”