May 1, 1913.; 



THE INDIA RUBBER WORLD 



413 



For this purpose we at first employed a dash pot, but later it 

 was found more satisfactory to use a device which prevents 

 recoil by holding the springs under the maximum tension. We 

 realize that this method of measuring the load is in a way open 

 to criticism, but at the same time we are gratified to find that 

 after more than two years of continuous service the corrections 

 for the dynamometers are so small as to be negligible, being well 

 within the limits of variation of the materials tested. We have 

 provided a set of weights with which the machines are calibrated 

 from time to time, and should the dynamometers become in- 

 accurate it will be a simple and very inexpensive matter to have 

 the old springs replaced by new ones. 



As already mentioned the measurement of elongation at rup- 

 ture is a less simple matter in the case of a straight test piece 

 than with a ring, the reason being that the rubber has a tendency 

 to creep to a greater or less e.xtent through the gripping surfaces, 

 thus making it inaccurate to take measurements between fixed 

 points on the grips. A little practice however enables one to 

 observe the extension between gauge marks on the test piece with 

 a fair degree of precision by pressing the end of a measuring 

 scale gently against one of the marks and taking the scale read- 

 ing just opposite the other mark as the specimen breaks. The 

 dynamometer automatically registers the breaking load. 



In my own experience the large amount of time consumed in 

 testing the elasticity or what is commonly termed "set'' or "re- 

 covery" after extension, has prompted me to devise means for 

 expediting the work, with the result that a piece of apparatus 

 has been constructed in which six specimens may be tested at 

 once. 



Of the tensile machines that I have had an opportunity to see 

 in operation, or to study from illustrations, some appear to be 

 rather better adapted to experimental or research work than to 

 the requirements of routine testing. This is due partly to deli- 

 cate or expensive construction, and also to the fact that more 

 time is required in their operation than is desirable in ordinary 

 routine work, for w'hich, simplicity of construction, ease and 

 rapidity of operation, would seem to be very desirable to say the 

 least. For reasons already stated it would be difficult to obtain 

 by autographic means, tension or hysteresis curves for straight 

 test pieces. A rather slow but otherwise satisfactory method of 

 securing curves may be carried out in the following way: 



A narrow flexible scale either of paper or tracing cloth is 

 attached to the test piece at the upper gauge mark by means of a 

 small spring wire clip. From this scale one operator reads and 

 announces the elongation at convenient intervals of 25 to 50 per 

 cent., to a second operator who observes and records the corre- 

 sponding tension. With a little practice this process becomes 

 exceedingly simple and the points when plotted are found to give 

 very smooth curves which clearly show the characteristics of 

 different rubbers. Such curves offer a convenient means of 

 studying the tensile properties of rubber when cut in different 

 directions : and when compared with similar curves obtained 

 with ring specimens the influence of the form of test piece is 

 clearly apparent. In testing twenty or more different com- 

 pounds I have found that specimens cut longitudinally or in the 

 direction in which the sheet has been rolled through the calen- 

 der give curves that lie well above the corresponding curves for 

 specimens cut in the opposite or transverse direction. Ring 

 test pieces produce curves which lie below those for longitudinal 

 specimens and which sometimes coincide very nearly with the 

 curves for transverse specimens. 



The tension test is generally considered to be the most satis- 

 factory and most widely applicable test for soft rubber products. 

 In its various modified forms it is used to determine the more 

 important physical properties such as strength, ultimate elonga- 

 tion and elasticity of recovery after a definite extension or a 

 definite load. The methods followed in making these determina- 

 tions vary according to the judgment of different persons, and 



since the results obtained are thereby influenced to a marked 

 extent, the importance of uniformity of methods is not open to 

 question. Of the influences that appreciably affect the results of 

 tests may be mentioned the following, which being famihar to 

 those interested in rubber testing would make an interesting 

 subject for open discussion: 



1. Shape and size test pieces. 



2. Method of preparing and measuring test pieces. 



3. Design of grip. 



4. Direction in which test pieces are cut (when straight 

 specimens are used). 



5. Temperature. 



6. Previous stretching. 



7. Time allowed after extension and release, before measur- 

 ing set. 



8. Speed at which rubber is stretched. 



Reference has already been made to the influence of the form 

 of test piece, and as regards the area of cross-section it appears 

 that in the case of rubber, as in the case of other materials, there 

 is a tendency for small test pieces to develop higher unit values 

 for strength and ultimate elongation, than larger ones. 



2. The preparation of test pieces and the measurement of 

 cross-section is generally conceded to be of the greatest im- 

 portance, and unless this part of the work is carefully and ac- 

 curately performed, it is useless to expect uniform results to 

 follow. When one considers that the width of ring as commonly 

 used is only 4 mm it is seen that a very slight eccentricity of say 

 Vi mm introduces an error of 12^/2 per cent, in strength. In the 

 case of straight specimens this particular difficulty does not pre- 

 sent itself, but here also it is necessary to make sure that the 

 cutting edges of the die are parallel and sharp. The usual 

 method of cutting straight specimens I believe, is to strike the 

 die a sharp blow with a mallet, but when used in this way, 

 unless the blow is well directed, there is a tendency for the die 

 to drift to one side, thus producing a specimen whose cross- 

 section is not rectangular. It would therefore seem worth while 

 to employ mechanical means to insure that all specimens are cut 

 alike. In our work we have found that an arbor press is per- 

 fectly satisfactory, and we use a piece of leather under the rub- 

 ber to prevent injury to the cutting edges of the die. 



For measuring the thickness, a spring micrometer is probably 

 the most convenient and accurate instrument available. 



3. The design of grip is of importance in that excessive local 

 pressure often causes failure at that point. For this reason it 

 would seem desirable to use a form of grip which closes auto- 

 matically as the tension is increased, and produces a uniform 

 pressure across the end of the test piece. 



4. As is generally known the tensile properties of a straight 

 test piece are influenced by the direction in which it is cut and 

 it is unfortunate that specifications sometimes fail to take this 

 fact into consideration. For example, a common specification 

 for water hose in this country limits the set after 300 per cent, 

 extension, to 25 per cent, and a rubber compound used for such 

 hose might show a set measured in the longitudinal and trans- 

 verse directions of. say, 28 and 22 per cent., respectively. Unless 

 this point were definitely covered therefore the hose might be 

 accepted or rejected according to direction in which the test 

 pieces were cut. The same is also true though to a less extent, 

 in the case of tensile strength and ultimate elongation. This of 

 course does not apply to ring specimens in which tension is 

 exerted equally in all directions. 



5. The influence of temperature on set, strength and ultimate 

 elongation, though often disregarded, is undoubtedly great 

 enough at times to justify consideration. When one considers 

 that the temperature in a testing laboratory may easily vary 

 twenty degrees or more between summer and winter it is not 

 difficult to account for variations in results which might other- 

 wise be overlooked. 



