Machine for Testing Rubber Products Used to Absorb Vibration'

one way or another. Nevertheless, no universally accepted laboratory tests, of significance as far as actual service con- ditions are concerned, have ...
0 downloads 0 Views 675KB Size
August, 1928

INDUSTRIAL AND ENGINEERING CHEMISTRY

853

Machine for Testing Rubber Products Used to Absorb Vibration’ Franz D. Abbott THEFIRESTONE TIRE& RUBBERCOMPANY, AKRON,OHIO

ITHIN the last two years numerous rubber parts, data nor comments are given. None of the above tests other than tires and tubes, have become widely simulate service conditions to any extent. Because of these unsatisfactory conditions a new machine, adopted in the assembly of automobiles. Virtually all of these new rubber parts are used to absorb vibration in called a compression flesometer, was designed, which tests one way or another. Nevertheless, no universally accepted the resistance of vulcanized rubber to flexure under comlaboratory tests, of significance as far as actual service con- pression. ditions are concerned, have been developed. Description of Flexometer The quality and type of rubber products for such service hare generally been determined from tensile, modulus, elongaThe compression flexometer (Figure 1) consists of a fixed tion, permanent set, and hardness data. It can easily be bottom plate with rocker holts hinged to it which pass through shown that such testJs not only are inadequate but may be a top plate, which is attached to a driving device. Bv screwing nuis down 0 ; t h e s e very misleading. Furtherrocker bolts the top plate more, most of them cannot may be made to produce a b e m a d e o n t h e finished A machine for studying the effect of flexing rubber compressive load on a test product. Although a stock under compression is described. The operation of piece placed between the p o s s e s s i n g h i g h tensile this machine In the testing of shock insulators, motor two plates. The top plate strength may be considered supports, and other rubber parts that absorb viis made to oscillate a t 800 as of better quality than one bration, or undergo flexure under compression, is cycles per minute. The ospossessing low tensile, in the given in detail. Ordinary tests such as tensile and c i l l a t i o n s are secured by case of rubber products used elongation are insufficient in evaluating such products. setting the thrust arms off to absorb vibration there Furthermore, the more recent static compression-set center with respect to the certainly is a value above test is not satisfactory. Results of tests on various which increase in tensile does driveshaft. It is possible to stocks as given with this device check with dynamomincrease the range of oscillnn o t m e a n better quality. eter and service tests. This is, of course, true of tory movement (throw) to Under the conditions of test adopted the permanent over 10 cm. (4inches) overmany other products also. set is shown to approach a maximum in 3 or 4 hours. Degree of cure has a deall by means of the slots cut The temperature of the test piece rapidly reaches a cided effect on the degree to in the eccentric collar shown maximum value. which vulcanized rubber at A. Tests r e c o r d e d i n Data and curves are presented for comparison of set this paper were made with withstands various service under flexure with tests under tension and with static a 0.95 cm. P/,-inch) over-all conditions. Nevertheless, compression. m a n y properties, particut h r o w . The compressive l o a d i s produced by the larly hardness, show n o decided change with relaweight of the top plate and tively large variations in cure. (“Hardness” of a rubber a component part of the thrust arms (a total of 13.6 kg. or product may be defined as the resistance which the surface 30 lbs.) plus that secured by tightening the nuts shown on the layers of this product offer to indentation by a penetrating top plate. A standard load is chosen for a given size of test piece. The tests herein reported were run a t constant distorelement under a predetermined load.) It is also to be noted that motor supports, rubber spring tion. For a purely technical study of rubber stocks, involvshackles, torque insulators, balls for universal joints, and simi- ing the effect of pigments and other ingredients, tests can lar parts, undergo permanent set due, to a greater or less also be made under a constant load by means of lead weights degree, to flexure under compression. fastened to the top plate. This work will be reported in a future paper. Note-In this paper any force applied toward the interior of the rubber The duration of the tests under various conditions, alsample will be called a “compressive load.” The deflection produced by such a force will, for the sake of simplicity, be called “compression” or “comthough arbitrarily chosen, is such that the permanent set pressibility,” I n the work on materials used t o absorb vibration we are developed in various high-quality stocks is great enough probably little concerned a t present, a t least, with the true compressibility to insure a low percentage error. Tests on materials to abof vulcanized rubber, since this is so small as to be negligible. sorb vibration are stopped far short of complete physical Although there have been static compression tests for per- breakdown or blowout, which is, in general, an explosive manent set under a definite load and also under a definite rupture of the test piece due to gaseous and liquid products distortion, there is a lack of uniformity in procedure and in produced by the heat of flexure. results. It is shown later in this paper that one of the most common of these tests fails to evaluate such products, and, Method of Making Test and Calculation of Results as far as the writer is aware, no Jatisfactory dynamic tests have been described in the literature. Mention is made of a The test piece, which has been molded, or cut and buffed, repeated compression test in an unsigned article,2but neither to the desired shape and size, is carefully centered on the bottom plate of the compression flexometer. Care is taken 1 Presented before the Division of Rubber Chemistry a t the 75th Meetthat flexure and loading occur in the same respective direcing of the American Chemical Society, St. Louis, M o , April 16 t o 19, 1928. 2 Rubber A g e , 20, 601 (1927). tions as in service. The top is then carefully lowered onto



Vol. 20, No. 8

INDUSTRIAL A N D ENGINEERIKG CHEMISTRY

851

from laborstory compression slabs 1.9 X 1.15 X 2.54 em. X Is/$ X 1 inch). (In giving the sizes of test pieces the width, length in the direction of flexure, and height, respectively. are stated in the order named.) Table I-Accuracy I'LBXURE-ssT

TIMEO. F ~ s x u n e DZ~~LBCTION Hrmrs Per'd

STOCZ

S ~ m ~ 1l e

Sample 2

8

36.0

36.0

2

55.5

3

34.4

34.4 31.2 41.7

2 2

30 60

2

50

30 1

29.3 43.1

Physical Results of Flexure under Compression

si*u,e 1--CampreSsion Pierornetel

the sample while the machine is a t dead center. The desired compression (uniformly secured) is obtained by means of the nuts on the top plste. The flexing is then started. Ordinarily the sample is removed from the flexoinetrr inunedistely at the conclusion of the flexure and allowed to cool 1 hour. Tho teat piece is then bisected in ti vertical plana in the dircction of flexure. The least height of this cross section in the direction of the loading is measured. The loss in height is reported as per cent flexurc-set. With stocks of t,he pure-gum type virtudly no set is developed under the ordinary conditions of test. In order to produce flexureset in such etocks the severity of- the test must be increased. This may be done hy increasiiig.tlie pressure, increasing the duration of the flexure, incrrasinrr the amplitude of flexure, or, by allowing the test piece

the tests.' Results of k t s run on steel plates with resulting lower ternperatures, due to conduction, will be reported in a subsequent paper, which will also include tests conducted without the top plete (zero ,comnrcsioni. ~~,~ , ~~

~

Accuracy of Test 'The accuracy obtained with this machine has been within per cent flexure-set. A large portion of the error involved is due to the fact that no effort has heen made t,o measnre the recovered height of the sample with an accuracy greater than 1 per cent. A good idea of the arcuaey can Le secured froni the data of Table 1. If the flexure-ret values of two different stocks check &s close as 3 per cent, tho tests should be repeated before concluding lhat one stock is better tlian the other. Thr tests recorded in Table 1 were made on test pieces cut -1

Examples of blow-out that occurred during Bexometer tests on portions of a solid tire are shown in Figure 2. These test pieces mrisured 5.7 x 4.4 X 6 em. (2'/* X 13/,X ZS/s inches) bphire the test. Failure occurred in a plane almost perpendicular to the compressive load (55.25 per cent deflection). Inspection showed a number of shorter slits pwallel to the maill opeliiiig. A11 evidence seems to indicate that failure had occllrred "along the grain" produced in the process of manufacture. Less severe tesds are desirable for shock insnlators and other motor products used to absorb vibration. Consequently much smaller test picces Were used. Under a low compression (10 to 15 per cent) less flexure-set results over a period of several days than is secured a t 50 per cent conipression in an hour or two. The action in a shock insulator and certain types of motor supports can be fimulated very well by proper choice of the load. Under these less severe conditions of flexure there a.re produced several residual stress= in the test pieces in addition to sub-permanent set measured after 1-hour recovery. Such effects can be seen in t,he buckled appearance the bisected test piece. Figure 3 shows a I-inch ball before test, (A), and after test (2 and 3). Sample 2 is a biser:t.cd ball of stock No. 3, and 3 is a bisected ball of stock No. 3. It, is easily seen that stock No. 2 undergoes less flexitre-set than does stock No. 3. Close inspection will show that the least, distance aero.% the cross section of 3 is considerably less thaii across a similar cross section taken farther from the

A-Compressive

lead; B-Direelion

01 deaure

Pisure I-EI~IZIDIPBof Blow-Out on Porrions of a Soiid Tire

center of the ball. This phenomenon was hardly noticeable in the case of stock No. 2. and is the best visual evidence of ~~~.~ . the stresses remainhig after flexure has ceased. Figure 4 shows a torque insulator (A), and a similar insulator (E) from which one ami has been cut for the test piece; also the test piece (C) before test, and flexed samples of stocks NO?.2, 3, and 16. There are also shown the directions of flexure ( Y Y ' ) and of the compressive load ( X ) . The line M N shows where the insulator is cut for the test piece. It is easily seen that stock No. 2 suffered less Resure-set than either No. 3 or 16, and also sives evidence of lms rcsidual stresses ("buckled" to a less degee). Figure 5 further illustrates the effects of flexure under compression. Stock No. 14, a so-called pure gum, sliowed

INDUSTRIAL A N D ENGINEERING CHEMISTRY

August, 1928

zero flexure-set. Stock 15 showed 45 per cent. This test piece suffered blow-nut, so WRS not bisected. Blow-out had relieved the stresses that would otherwise have remained after flexure. Table I1 gives general data pertaining to the types of test pieces shown in Figures 3 to 5, inclusive.

are given in Table IV. pression. 148" C. Mi"XiO* 65

75

83

Effect of Various Factors on Flexure-Set

TIME OF FLEXURE-Flexure-set rapidly approaches 3 maximum RS the test progresses. Temperature rises even more rapidly and probably inflnence? to a considerablr ex-

3

A

Figure 3-Effect

of Flexure of Rubber Ball

tent the slope of the flexure-set-time curve. This is seen by comparing curves I and 11, respectively, of Figure 6. The data for the temperature-rise curve (11) were obtained on the 210-minute test (Table 111). A copper-constantan t,herinocouple was imsert.ed in 3 narrow horizontal slit near the bottom of the test piece before its assembly in the flexometer. Qadings on t.he potentiometer were taken eyery few minutes for 2 honrs. The data for tests included in curve I are giren in Table 111. Flexure NBS a t 55.5 per cent compression on test pieces cut from two insulators of stock No. 1, cured 70 minutes

12.4 23.5 30.6

5

16 30

60 I20 210

Flexure was at 55.5 per cent com-

Table IV-ERecf CURC AT

70

Table 11-Test Data for Different Test Pieces

855

I Hour Per ran: 40.2 36.2 33.3 26.4

of Cure Frrixms-Snr

2 Isours PE" e n *

43.1 43.1 43.1 33.3

H n ~ n ~ ~ s s - H a r d n eis s s no indication of the resistance of a stock to flexure under compression, as shown by the data in Table V. Motor part 1had been cured to the proper state of cure as determined by service and by flexure-set data. Motor part 2, of a different size but similar shape, was also cured to the proper hardness values. The test pieces were flexed for 2 hours under 55.5 per cent distortion. X ll/s X All test pieces were 1.43 X 3.8 X 2.86 em. I'/S inches). I n every case motor part 2 showed excessive flexure-set. These data show the fallacy of using hardness data to indicate that cure which gives the greatest resistance to dynamic fatigue,

COMPRESSIRSLITY (DEFI.ECTroX)-~atR on the deflection of 3 test piece of standard size by a given compressive load fail to furnish a satisfactory indication of resistance to flexure. (Table VI) In these tests disks 1.9 em. ( 3 / r inch) in diameter were cut from 3 slab of 0.fi:i-cm. ('/$-inch) gage. These were then subjected to a load of 42.18 kg. per sq. ern. (600 pounds per square inch). Stock No. 4 shows the greatest resistance to flexure, yet it stands ahoiit midway in order of compressibility (column 2, Table VI). On the contrary, stock No. 7 shows the greatest deflection but stands midway in order of resistance to flrxure under compression. STATIC-SET A N D IIYSTERESIs LOSS lJ?.7lEJ< collPllESSSON-

Permnnent set under simple compression does not clieck with flexure-set,, whereas hysteresis loss does give n quite

36.2 43.1 44.4

STATE OF CURE-Decided undercure results in a high flexure-set. As the time of enre progresses the resistance to flexingunder compression increases. Flexureset approaches 3 minimum. as evidenced by the fact that overcures were fonnd to show high flexure-set again. Curve I, Figure A 16 2 3 C B 7, shows that flexure-set decreases as the state of Figure .(-Torque Insulstorand Flexed Samples of Stocks Nos. 2. 3. and 16 cure nroaesses. The tests were of 1-hour duration on test pieces from shock insulators cured 65, 70, 75, and satisfactory indication of resistance to dynamic fatigue. 80 minutes at 148" C. (stock No. 1) Cume I1 shows flexure- Some data comparing these tests :ire giren in Table VII. set resulting from 2 hours' flexure for the same cures. The 2- Ench stock is seen to stand in the same relation to t.he others, hour test is loo severe to differentiate bet.ween the lower cures. whether flexure-set or hysteresis data are considered. Staticbut does show the 80-minute cure to be decidedly better than set (permanent set under static compressive conditions) the ot.hers. Nevertheless, it has the advantage that all prod- data are far out of line. The lzystereiis vi~lueswere deteructs showing a greater flexure-set than 3 "standard" adopted mined by getting tho load-deflcvtion dntn on circular disks value bnsed on t.his duration are eliminated and a high- 1.9 cm. in diameter and OS!:< m . thick. The maximum quality product is thus assured. The data for these tests load was 42.18 kg. per sq. en). 'I I:c initinl deflection a t this .

Y

INDUSTRIAL A N D ENGINEERING CHEMISTRY

856

13

14

15

11

Figure 5-Effect of Flexure of Laboratory Samples

load is given in column 5, Table VII. (A full report of this work together with a description of the instrument designed to study hysteresis loss and permanentset under either constant deflection or a constant load, will be given in a subsequent paper.) Static-set tests were made on similar disks under the distortion secured by a load of 42.18 kg. per sq. em. Tests were conducted in an oven at 82' C. for 24 hours. Table V1-Compressibillty

1

14.5

30

I1

14.2 26.0

21.0

12 14

14.8

3 10

15

8.9

..

10.0 11.0 0 45.0

78.

Vol. 20,KO, 8

It will be noted that the values for ultimate elongation are alike within experimental error (with the possible exception of stock No. 8). Set at break (the increased distance between two marks originally l inch apart, measured 10 minutes after the dumb-bell strip has broken) likewise fails in evaluating these stocks. Ultimate tensile likewise is of little value for this purpose. It is to be noted that stock No. 3 possesses the highest tensile (217 kg. per sq. em.) but suffers a higher flexure set (19.4 per cent) than does stock No. 4, which has an ultimate tensile of only 166 kg. per sq. cui. At the other extreme is stock No. 6 with the lowest ultimate tensile (140 kg. per sq. em.), which also shows a still higher flexure-set. Similar discrepancies are to be noted in the case of moduli. Stock No. 3 with the second highest stress at 400 per cent elongation (116 kg. per sq. em.) and stock No. 6 with the lowest stress at 400 per cent elongation both suffered greater flexure-set than stock No. 4, which possessed an intermediate value for the stress at 400 per cent elongation and also the lowest flexure-set. Table I X shows to much better advantage how poorly data of tests under tension evaluate the above stocks as to their resistance to dynamic fatigue.

Flexure-Set

27.8

5.85 4.8

3:95 3.00 10.53

P'L)

4

33.7 37.3 51.8 43.0 45.0

..

REPEATED-STRETCH SET-Set a f t e r repealed e l o n g a t i o n '?do% does not help in the ili evaluation of s t o c k s 3 as to their resistance to flexure under compression. It was b o lieved tha,t tliis test, lo%61 70 75 80 being a dynamic one, CURe IN M ~ W T E S althoiigh under ten, I%*ure7 sion instead of compression, might properly evaluate stocks. This certainly should be the case if w e believe t,hc same as do Gottlobs and othem4 The data in Table X show conclusively that set after repeated stretch under either constant load or constant distortion does not check set resulting from repeated flexure under compression.

3

Table X.-Sef Due to Repeated Stretch YS. Plerure-Set SET DUBTO R B P E A T CSTRBTCX ~ constant constant n'exunaSTOCP

load

distortid

SeT

r e * canl 20.0 21.9 10.0 11.0 37.5 45.0

Figure 6

FLEXURE-SET T'S. TENslLE STRESSES, ULTIMATE ELONQASET AT Da~.4~-The ordinary data from tests under tension also fail to evaluate satisiactorily the resistance of stocks to flexure under compression. Comparisons of some of lhese data with flexureaet are &\en in Table YIII. TION, A N D

Broke duiinz first rlonrdon.

b Broke during secoud elongalion

Determination of Stretch under Constanl Load-.% dumb-hell test strip was stretched five times in a tensile machine at 51 cm. per sec. (20 inches per minute) to 70.3 kg. per sa. cm. (1000 pounds per square inch). The interval between extensions was only that necessary for the machine to reverse. After the fifth returii the strip was removed and allowed to recover for 1 hour. Permanent set was measured as the increased distance between two marks oiigiiially 2.54 em. (1 inch) apart. ~

a 4

"Technology of Riibhci." PP. 199 and 202. English ed.. lS27. Devien and A-orton, India Rubber J.. 69, 833 (1920).

August, 1928

I N D U S T R I A L A N D ENGINEERING CHEMISTRY

FLEXURE-SET 17s. DYNAMOMETER TESTS-Flexure-set sults check dynamometer tests very satisfactorily.

re-

Dynamometer Tests-Sixteen balls are assembled in a universal drive.unit, each under approximately 21.65 kg. (50 pounds) compression. The alternate balls seem to carry practically the whole load of running at 1500 r. p. m. under a torque of 8.9 kg.meters (64.5 foot-pounds). The generator shaft is at a 5-degree angle to the drive shaft. The test is stopped when the drive unit becomes loose.

It will be noted from the data in Table XI that stock No. 2 in each instance shows a greater resistance to fatigue resulting from flexure under compression than does stock No. 3. These data are typical of the results with other stocks. Table XI-Dynamometer Test vs. Flexure-Set FLEXURE-SET

DYNAMOMETER TEST Ball for Condition universal STOCKDuration of ball joint Hours Per cent 3 26a High set, 37.5 very poor 2 321/r K o s e t , bad 15.6 38’/rb a At 3.5-degree angle of drive. b At 5-degree angle of drive.

Torque insulator Per cent 31.8 20.4

1.9-cm. compression slabs Per cent 30 12

857

The fact that flexure-set results for stock No. 2 (or No. 3 as the case may be) do not check within themselves does not discredit the flexometer test. On the contrary, the fact that two stocks line up in the same order although tests were conducted on different volumes and shapes of test pieces as well as under different deflections and different durations of test (Table 11), seems an important point in favor of the flexometer test. Summary

Data have been presented to show that the commonly used physical tests such as stress a t break, stress at 400 per cent elongation (modulus), ultimate elongation, set a t break, set after repeated stretch, and static compression-set fail to evaluate properly the resistance of vulcanized rubber to dynamic fatigue of flexure under compression, whereas flexureset as determined in a new machine (the compression flexometer) herein described does evaluate stocks properly as determined by service and dynamometer tests. Accordingly, resistance to flexure under compression (dynamic fatigue) is suggested as a much more important property of rubber products used to absorb vibration under compression than any of the above-mentioned tests.

Jelly-Strength Measurements of Fruit Jellies by the Bloom Gelometer’ Carl R. Fellers and Francis P. Griffiths DEPARTMEKT OF HORTICULTURAL MANUFACTURERS, IV~ASSACHUSETTS AGRICULTURAL COLLEGEAND

EXPERIMENT STATION, AMHERST,MASS.

With a few minor modifications the Bloom gelometer evaluating the jelly-making was found to be very satisfactory for the determination ment for giving reaquality of fruit juices and of the jelly strength of fruit jellies and similar prodsonably accurate and commercial pectins. A deucts. Jellies may thus be standardized as to conreproducible measurements scription of its adaptation and sistency within relatively narrow limits. This is of of the jelly strength of jellies use for the measurement of significance in t h e manufacture of jellies has long been r e c o g n i ~ e d . ~ , ~ particular ~~ t h e j e l l y strength of fruit and similar products as well as in the grading of comBogue3 described the several jellies should prove of value mercial pectins. types of instruments in use to both laboratory workers Jellies for testing should be sealed, stored for a t prior to 1922 for testing gelaand manufacturers. least 24 hours, and tested a t a constant temperature, tin and g l u e p r o d u c t s . The Bloom gelometer is preferably from 20” to 23“ C. Weak jellies set more S ~ c h a r i p a ,also ~ Baker5 and already a standard instruslowly than firm ones. The greater part of the change described instruments ment for the measurement of occurred during the first 24 hours. Jelly strengths involving the use of water or edible gelt~tin,~ this product decreased with rise in temperature. Owing to skin air pressure for the testing of being usually sold on the basis effect on the surface, uncovered jellies gave unreliable pectin jellies. Tarr s t a t e d of jelly strength as measured readings. that he found no inRtrument b y t h i s instrument. The Considering both top and bottom readings the used for the measurement of latter may be purchased togelatin or glue adaptable to average per cent deviation of twenty-two series of gether with the chill bath repectin jellies. samples, ranging from very weak to very strong, was quired for gelatin testing from The Bloom gelometer’ as approximately 5 per cent. the secretary of the Edible originally described by RichJellies of good organoleptic qualities, consistency, Gelatin Manufacturers’ Reardson* has been used in our and texture usually tested between 60 and 100 grams search Society of America. laboratory for two years for and strained cranberry sauce tested between 140 and Operation of Instrument making jelly strength deter180 grams. m i n a t i o n s on fruit jellies, Three m o d i f i c a t i o n s i n jams, and similar products. It has also proretl useful in operation were found necessairv in order to adaDt the instrument to the testing of fruit jellies: (1) The distakce of travel 1 Received M a y 2, 1928. Contribution h-o. 78, Massachusetts Agriof the plunger was increased to 5 mm. (2) The heavier metal cultural Experiment Station. 2 Paine, A m . Food J., 17, No. 3, 11 (1922). cup receiving the *hot was replaced by a light paper cup. “Chemistry and Technology of Gelatin and Glue,” p. Z69, McGraw(3) The rate of inflow of the shot was adjusted to 100 grams Hill Book Co., New York, 1922. in 10 seconds when the lever arm E1 was on the second dog from 4 “Die Pektinstoffe,” p. 81, Serger & Hempel, Braunschweig, 1925. the bottom. (Figure 1) Ih-D. ENG. CHEM., 18, 89 (1926).

HE need of an instru-

T

6 7 8

Del. Agr. Expt. Sta., Bull. 142 (1926). E. S. Patent 1,540,979 (1925). Chem. Met. E n g . , 28, 551 (1923).

0 Edible Gelatin Manufacturers’ Research Society of America, “Standard Methods for Determining Viscosity and Jelly Strength of Gelatin,” 1457 Broadway, New York, N. Y .