SLIP CONSTANT FROM 20 TO 40 MIN.

In view of this fact, a careful study was made in an endeavor to find the cause for the variation. By taking continuous observations of slip over long intervals of time, with the machine running under constant conditions of tension, brake load and speed, data was collected and curves plotted between per cent. of slip and time. Those on Figure 3 are typical. From these it will be seen that the slip is continually varying. By comparing a large number of such curves it was found that this variation takes place through no clearly defined cycle, although there is in general a tendency for the slip to be considerably higher when the brake load is first applied than subsequently, over even a short interval of time. This condition also probably explains some of the peculiar speed variation noted at times by operators of machines driven by cast iron pulleys when the speed of the driving shaft remains apparently constant. It also accounts in a large measure for the wide variation in experimental data and results.

Referring to the upper curve of Figure 3, it will be seen that if the observations of slip had covered only the first five minutes of the test, the average would be about 10 per cent., while if of the same duration but beginning fifteen minutes after the test started, the average would be only about one-half of 1 per cent. If taken twenty-five minutes after the test started, the result would be a little over 5 per cent. In the second curve on this plate the variation is not quite so extreme. Yet these curves well illustrate the fact that a few average observations, though conscientiously taken and covering what might be considered a reasonable length of time are almost certain to be misleading. This led to a determination to take a sufficient number of observations in these tests, and over a great enough duration of time to insure obtaining at least, approximately average results.

THE TESTING MACHINE.

The machine used in these tests was that designed by Mr. WILFRED LEWIS, and built by the William Sellers Company of Philadelphia in about 1885, which has been in the laboratory of

Sibley College for a number of years. In order to simplify the apparatus, and as drive efficiency was not sought, the transmission dynamometer was removed. Several important improvements have also been made in the way of a better slip meter, and a new arc of contact meter has been added. There have also been made some changes in the belt tension weighing device.

The machine consists of the following principal elements: the driving shaft, a tension weighing device, an arc of contact meter, a slip meter and a driven shaft with Prony brake and accessories mounted on a movable carriage.

Figure 4 shows the arrangement in detail: (A) is the main driving pulley; (B) a universal coupling; (C) the tension weighing frame which is carried at one end by two thin steel plates mounted horizontally at c-c and at the other end by the scale (D); (F) is the arc of contact meter and (G) the slip meter, consisting of a graduated disc (Q) keyed to the shaft (H), and a pointer (k) attached to the hub of the worm wheel (K). Worm wheel (K) is driven by a worm on the end of the shaft of the driven pulley (M) and is free to turn on its own shaft (H); while worm wheel (L) which is keyed to this shaft

(H), is driven by a worm on lay shaft (S), driven in turn by gears of equal ratio from the driving shaft (T). (N) is the driving and (M) the driven test pulley; (O) the Prony brake and (P) the scale upon which its loans are measured. The tension in the belt is adjusted by means of a lever (R) which slides the carriage with the driven pulley along a track by means of a pinion engaging a rack attached to the floor. At the rear of the carriage a snub bar is attached to hold it in any given. position.

OBSERVATIONS.

The following observations were taken: Initial tension, tension when running light, running tension, brake load, arc of contact and slip.

The initial tension was taken at the beginning of each test for four positions of the driving shaft, 90 degrees apart and averaged. The machine was then started, brought up to speed and the tension taken while running light. A brake load of 20 pounds was then applied and the running tension, arc of contact and slip taken requiring a minimum of about four minutes time. The running tension was then again measured, after which a next higher load was applied and the same observations repeated. This was continued until the maximum load which the belt would carry, or which could be safely put on the machine, was reached. The machine was then stopped and the final tension taken in the same manner as the initial tension.

The brake load was kept constant during each set of observations by an operator detailed for that purpose. The speed of the driving shaft, which was controlled by a rheostat in the field of the driving motor in conjunction with an armature rheostat, was maintained constant by a second operator, having a Horn tachometer attached to the shaft of the driving pulley.

REDUCTION OF OBSERVATIONS TO RESULTS.

From the tests the observed data for each initial tension was then plotted, using as co-ordinates the net brake load and the per cent. of belt slip, and a curve representing a mean of the observations drawn.

This method is illustrated by Figure 5, which shows readings plotted for an initial belt tension of 150 pounds per square inch section. The readings are in each case indicated with the number of the test in which they were obtained. None of these were established by less than ten observations and in some cases there were as many as eighty observations. Thus at ninety pounds brake load in test No. 10, a series of ten consecutive readings each gave about 3 per cent. slip, while the average of fifty readings gave 12.3 per cent. The broken line curve shown in Figure 5 was chosen as representing the average relation for this tension. In determining on this curve, however, we were as much influenced by the relation which it should bear to the curves for the other tensions as by any other single consideration.

Figure 6 shows the entire group of curves for the six tensions used. From this figure the values of slip for the various brake loads were taken and recorded in column 5 of the tabulated results shown in Table 3. Owing to the fact that the slip meter used in these tests gave the slip in terms of the driven pulley speed and it is customary to express such relations in terms of