These experiments seemed to show that the principal resistance to straight belts was journal friction, except at very high speeds, when the resistance of the air began to be felt. The resistance from stiffness of belt was not apparent, and no marked difference could be detected in the power required to run a wide double belt or a narrow light one for the same tension at moderate speeds. With crossed and quarter-twist belts the friction of the belt upon itself or upon the pulley in leaving it was frequently an item of more importance, as was shown by special experiments for that purpose.

In connection with the experiments upon internal resistances, some interesting points were noted. Changes in tension were made while the belt was running, commencing with a very slack belt and increasing by definite amounts to the working strength. As this point was approached, it was found necessary, to maintain a constant tension, that the tightening bolt should be constantly operated on account of stretch in the belt. Then, again, as the tension was reduced from this limit, it was found that at lower tensions the belt would begin to shrink and tighten for a fixed position of the sliding frame. This stretching and tightening would continue for a long time, the tightening being, of course, limited, but the stretching indefinite and unlimited.

The first series of experiments was made upon paper-coated pulleys 20′′ diameter, which carried an old 51⁄2′′ open belt 3⁄16′′ to 1⁄4′′ thick and 34 ft. long, weighing 16 lbs. The arc of contact on the pulleys has been calculated approximately from the tension on slack side, and for this purpose the width and length of the belt were taken. The percentage of slip must be considered as equally divided between the two pulleys, and from observations made it is easy to calculate the velocity of sliding when the speed is given.

Some of the most important results obtained with this belt are given in [Table I.] in which the experiments have been selected to avoid unnecessary repetition. In all cases the coefficient of friction is shown to increase with the percentage of slip. The adhesion on the paper-covered pulleys appears to be greater than on the cast-iron surfaces, but this difference may possibly have been due to some change in the condition of the belt surfaces.

After a fresh application of the belt dressing known as “Beltilene,” the results obtained are even higher on cast iron than on paper surfaces, but after a time it was found that the adhesive property of this substance became sensibly less and less. Flakes of a tarry nature rolled up from the belt surface and deposited, themselves on the pulleys, or scaled off.

So much was found to depend upon the condition of the belt surface and the nature of the dressing used, that the necessity was felt for experiments upon some standard condition which could be easily realized and maintained. For this purpose a belt was taken from a planing machine when it had become perfectly dried by friction. The results of experiments upon this belt are given in [Table II]. When dry, as used on the planer, the coefficients for any given percentage of slip were much smaller than those given in [Table I]. This was naturally to be expected, and the experiments were continued to note the effect of a belt dressing in common use, known as “Sankey’s Life of Leather,” which was applied to the belt while running. At first, the adhesion was very much diminished, but it gradually increased as the lubricant became absorbed by the leather, and in a short time the coefficient of friction had reached the unprecedented figures of 1.44 and 1.37.

TABLE I.

STRAIGHT OPEN BELT 51⁄2′′ WIDE BY 7⁄32′′ THICK AND 34 FT. LONG, WEIGHING 16 LBS., IN GOOD PLIABLE CONDITION, WITH HAIR SIDE ON PULLEYS 20 IN. DIAM. RUNNING AT 160 R. P. M., OR ABOUT 800 FT. PER MINUTE.
No.
of
Exper-
i’nt.		Sum of Tensions T + t[41]
T - t
Work-
ing.		[41]
T		[41]
t		[41]
T ⁄ t		Per-
cent-
age
of
Slip.		Velocity
of Slip
in ft.
per
minute.		Arc
of
con-
tact.		Coef-
ficient
of
Friction.		Remarks.
Initial.Work-
ing.		Final.
17200210 100155 55 2.82 .41.6177°.336Paper-covered
pulleys.
19 220 140180 40 4.50 .62.4176 .490
21 246 180213 33 6.451.24.8175 .610
22 260 200230 30 7.672.610.4174 .671
23 270180220245 25 9.807.931.6173 .756
24300316 200258 58 4.45 .72.8177 .483
27 344 260302 42 7.201.04 176 .643
28 350 280315 35 9 1.87.2175 .719
29 364 300332 32 10.42.811.2175 .784
30 380260320350 30 11.75.522 175 .805
31400422 200211 111 1.90 .52 179 .205
33 440 280360 80 4.50 .83.2178 .484
35 470 360415 55 7.541.14.4177 .654
36 506 400453 53 8.542.18.4177 .694
37 520380420470 50 9.405 20 177 .725
60200205 80147.567.52.18 .52 178 .251Cast-iron surfaces.
61 210 100155 55 2.82 .93.6177 .336
62 215 120167.547.53.521.76.8177 .407
63 220 140180 40 4.503 12 176 .490
65 246180180213 33 6.4512 48 175 .610
66300300 120210 90 2.33 .52 179 .270
68 310 160235 75 3.13 .83.2179 .365
69 315 180247.567.53.671 4 178 .418
70 320 200260 60 4.331.76.8178 .472
71 325 220272.552.55.192.610.4177 .545
72 340 240290 50 5.803.815.2177 .569
73 350 260305 45 6.775.522 176 .623
74 360 280320 40 8 8.634.4176 .677
75 375 300337.537.59 15.260.8175 .719
76400420 200310 110 2.82 .62.4179 .336Belt dressed with
“Beltilene.”
78 460 280370 90 4.111 4 179 .452
81 480 340410 70 5.861.56 178 .569
84 510 400455 55 8.272.28.8177 .684
86 535 440487.547.510.24.518 177 .760
88 560385480520 40 13 8.433.6176 .834
89300320 120220 100 2.20 .41.6179 .252
93 350 200275 75 3.67 .83.2178 .418
97 390 280335 55 6 1.66.4177 .580
101 440 360400 40 10 3.112.4176 .750
104 470310420445 25 17.88.634.4173 .953

[41] T represents the tension on the tight part, and t on the sag part of the belt.