A Study Of Different Types Of Steel For Grinding Media At Climax

In the course of the grinding experiments conducted at Climax over the past twenty years one interesting phenomenon was discovered. This was that balls made of certain types of steel grind more ore than some others. The most significant variation was found in the results secured with a hard alloyed forged bail compared with the forged carbon ball supplied by most ball manufacturers These findings were confirmed in many repeated controlled tests. The value of these tests depends on the methods used, the facilities available and the care with which the investigations were conducted. These are described in the following. This testing has largely been done in eight 91 x 81 low discharge ball mills in closed circuit with 7811 duplex screw classifiers. The mills run at 20 RPM on a feed approximately minus 3/8 inch. The ore is ground through 28 - 35 mesh in one stage. Each mill and auxiliary classifier is in a separate, parallel grinding and flotation section The, ore to each: section is the same in every particular and is fed to each mill bin from a common conveyor tripper. The ore from the bin to the ball mill is weighed on a conveyor scale. The power input to each mill motor is determined with a watt hour meter which measures the power to the mill motor only. Both the scales and meters are checked periodically by standard methods and have been found to give consistent readings. Automatic samplers are employed on each section to obtain accurate measurement of the size reduction. A size analysis of the classifier overflow is made each shift and a composite size analysis is obtained from a weighted average of the various shift analyses. The mills are driven by 450 horsepower synchronous motors so that there is no speed variation and shell liner age is kept as nearly the same as possible in mills under test. The same set of operators is assigned to a mill under test as to the control mill. In all other particulars such as ball load, mill density, liner design, etc., the mills are operated in precisely the same manner. This is an ideal arrangement for plant testing. From observations of ball mill performance over the past twenty years, it is the opinion of the operating staff that variations of as little as one to two percent in mill capacities may be determined if sufficient operating hours are devoted, to a test. Relations between mill capacities, fineness of grinds and power consumptions have been established with a high degree of accuracy. In Table I are given the figures for one test. These show that while the grind with the hard ball was slightly coarser, the tons ground per hour was 5.4 percent greater with the soft ball, and 6.1 percent greater on the basis of minus 100 mesh produced. The power consumed per ton with the hard ball was 4.4 percent greater on a tons per hour basis. The soft ball was inferior in wearing qualities and steel consumption per ton of ore was 24 percent greater with it. In this test the mills were charged with the two different balls for a period of three months to allow the ball loads to come to equilibrium. In Table II are given the specifications of the two balls. The difference in grinding rates may be accounted for by the nature of the ball surfaces. The surfaces of the soft balls after being worn a while are rougher to the touch. Enlarged photographs of the worn-in surfaces of hard and soft balls are shown in Figure I. These show the presence of scratches or skid marks on the hard ball surface which indicates less nipping effect. The Climax ore pulp has a pH of 6.0 and is in a dispersed state. The viscosity is low compared with most ores especially those that are conditioned with lime. This would lower the tendency of the pulp to adhere to the balls with a lower grinding efficiency as a result. It was also thought that the relative rates of wear from surface to center might affect the make-up of an equilibrium ball charge. This probably has little effect on capacity as shown by the size analyses of the two ball charges in Table III. This hypothesis was subsequently investigated by the use of rationed charges. The rationed charge did not overcome the advantage in