Selection of Volume of Ball Loads in Continuous Dry Open Circuit Grinding: The tests reported in tables 19 and 20 were made in cooperation with a large cement-manufacturing company that wished to obtain more grinding for the power expended. The ball load in its clinker- grinding mills was 18 percent of the mill volume and the company did not deem it consistent to increase the ball load because the mills, having been pebble mills originally, could not carry a heavier load; also, the ball loads already taxed the motors to their limit. The tests required investigating the ordinary overflow type of mill and the low pulp-level type. The tests shown in table 19 anticipated the overflow mill and those analyzed in table 20 were expected to illustrate the deportment of the low-pulp-level or quick-discharge mill. All were dry-feed tests. In table 19 the ball volume was from 18 to 45 percent, and the feed rate had to be regulated so that the subsieve sizes in the products were about the same.

No noteworthy difference in type of grind resulted. The recording of capacities is omitted because, obviously, capacity would be low when ball volume was low. The change in efficiency, however, is important. The efficiency of the overflow mill slumped badly when the ball volume was small. This is because the small ball volume left too much space for ore in storage. These variations are shown in the next to the last line in table 19, in which is given as nearly correct an estimation of the amount of ore in the mill as could be determined conveniently. The best work was done with 40- and 45-percent ball volumes, the discharge being 4 inches in diameter.

The last line in table 19 shows the ratio of weight of ore to weight of balls. This weight relationship is significant because it leads to the conclusion that when the ball volume is small too much power is spent tumbling the ore; with 18 percent ball volume more than one-third of the weight of the mixture was ore. Then, also, there is the question whether the action of balls would be orderly in such a large volume of ore—that is, would not the mill be little more than a tumbling device to mix ore and balls? Probably disorderly tumbling would result.

The importance of a correct ratio of amount of ore to balls is borne out by examination of the results of a test, shown in table 20, in which the grinding had to be in batch so that the ratio of ore to balls could be kept the same for the different ball volumes. When ratios were correct, the 18-percent ball volume performed nicely, although, for reasons already given, it had failed in the open circuit. The type of grinding and the efficiencies were about the same in the two tests. The conclusion was reached that cement-grinding mills with 18 percent ball volume were doing poorly if they were of the overflow type. If they were of the low-pulp-level type, and the low-pulp-level devices keep the clinker charge low in accord with manufacturers’ claims, the efficiency was not impaired because of the small volume of balls.