Tumbling Mill Characteristics
Tumbling Mill Characteristics
Nelson, J.E., "Mechanical-Electrical Coordination in Grinding Mill Scheibe, W., Dallmann, W., and Rosenbaum. A,, "Selection of Grind-
Design," Canadian Mining and Metallurgical Bulletin. Vol. 65. ing Additives." Silikattechnik Vol. 21, No. I, Jan. 1970, pp. I I-
No. 724, Aug. 1972, pp. 52-57. 17.
Olds, A.R., Jr., and Ho, T.. "Primer on Large Ball Mill Drives, 1, Schuler, U., and Hess, H., "High Throughout Bowl Mills," Brenns-
2." Minerah Proessing. Vol. 5, No. 3. Mar. 1964, pp. 36-39; toff-Waerme-Km/I. Vol. 20, No. 6, June 1968, pp. 278-283.
No. 4, Apr. 1964, pp. 38-41. Smith, A.R., and King, H.G., "Four Years Operating Experience
Prescott, T.W.L., and Webb, F.C., "Size Distribution Produced in with the Aerofall Mill at British Industrial Sand Limited, Money-
a Hammer Mill," TmIISaction~,Institution of Chemical Engineers, stone, United Kingdom." Proeedings, 8th Commonwealth Min-
Vol. 50, No. I, Jan. 1972, pp. 21-25. ing and Metallurgical Congress, Vol. 6, Melbourne, Australia,
Ramanujam, M., and Venkateswarlu, D., "Studies in Fluid Energy 1965. p p 611-617.
Grinding," Powder Technology. Vol. 3, No. 2, Dec. 1969, pp. Spencer, A.G., "Electrical Drives for Large Grinding Mills," Cana-
92-101. dian Mining and Metallurgical Bulletin, Vol. 65, No. 726, Oct.
Rathburn. D.R., and Mok, J.K.. "Fine Dry Grinding of Iron Ore 1972, pp. 73-79.
for Pelletizing." T m n s SME-AIME. Vol. 247, 1970, pp. 335- Tanaka, T., "Tube Mill Problems," Zement-Kalk-Gips, Vol. 10, No.
339. 12, Dec. 1957, pp. 530-533.
Rauth, A.A.. "Maintaining an Optimum Grinding Charge," Trans. Tartaron. F.X., "A General Theory of Comminution." T m n s SME-
SME-AIME, Vol. 244, 1969, pp. 82-88. AIME, Vol. 226, 1963, pp. 183-190.
Rich, E.A.E., "Concepts of Gearless Ball-Mill Drives," IEEE Tmns- Thomas. W.N., and Dickman, G.J.. "Selection of Drives for Large
actions on Industry and Geneml Applications, Vol. 5. No. 1, Jan.- Grinding Mills," T m m SME-AIME, Vol. 250. 1971, pp. 19-30.
Feb. 1969, pp. 13-17. Turner, H.E., and McCarthy, H.E., "A Fundamental Analysis of
Rose, H.E., "Gear Failures of Ball, Tube, and Rod Mills," Engineer, Slurry Grinding," AIChE Journal, Vol. 12. No. 4, July 1966,
Vol. 203, No. 5280, Apr. 1957, pp. 522-524. pp. 784789.
Rowland, C.A., Jr.. and Nealey, R.C., "Application of Dry Grinding Veitsman. L.Ya., et al., "High Output Attrition Mills," Thermal Engi-
Rod Mills." T m n s SME-AIME, Voi. 244, 1969, pp. 229-236. neering (USSR), Vol. 14, No. 8, Aug. 1967, pp. 125-127.
Rowland, C.A., Jr., "Applying Large Grinding Mills," Canadian Min- Wartenweiler. F.. "The Use of Composite Loads in Tube Milling,"
ing Journal, Vol. 91. No. 11. Nov. 1970, pp. 51-54. Journal, Chemical, Metallurgical, & Mining Society of South Af-
Rule, W.T., "Dry Autogenous Grinding at Butler Taconite," 29th rica, Vol. 32, 1932.
Annual Mining Symposium, University of Minnesota, M i n n e a p Weston, D., "Theoretical Concepts of the Aerofall Mill with Illustra-
lis, 1968, pp. 205-210. tions of Plant Practice," Proceedings, International Minerals Pro-
Rumpf, H., "Principals of Impact Comminution and Their Applica- cessing Congress, London, 1960, pp. 24-34.
tion in Jet Grinding," Chemie-Ingenieur-Technik, Vol. 32, No. White. H.A.. "The Theory of Tube Milling," Journal, Chemical.
3, Mar. 1960, pp. 129-135. Metallurgical, and Mining Society of South Africa, Vol. 15, No.
Rumpf, H., Symposium Zerkleinern, Symposium on Size Reduction, 7, 1915.
Frankfort, W. Germany, 1962. Zimmerman. G.T., "Ball. Pebble & Rod Mill Installations." Cost
Salman, T., and Sirois, L.L., "Density, Viscosity and Surface Area Engineering, Vol. 4, No. 4, Oct. 1959, pp. 4-16.
Relationship in Grinding," Canadian Mining Journal, Vol. 83,
No. I, Jan. 1962, pp. 3438.
rate or efficiency when the cones are used instead of balls in the while the most popular size was 3-in. diam. At present there is very
mills, though no conclusive data to support this claim have yet been little demand for the 4- and 5-in. sizes (except in semiautogenous
reported. mills) and the most popular size is now 2-in. diam.
During the period between World Wars I and 11, when many Matenah for Grinding Rods. Grinding rods range in size from
large milling operations to treat low-grade ores were built, it became 1% to Cin. diam and are normally 10 to 20 ft long. depending on
necessary to develop sources for the low-cost white iron balls which the length of the mill. Their size and shape practically dictate that
were extensively used during that period. New foundries, equipped they be made from rolled steel bars. The rods are cut to their specified
with specially designed metal-mold casting machines, were built at length, usually with a cut-off wheel which produces a square cut.
the milling operation or at locations where they could conveniently They are machine-straightened, usually to a tolerance of %-in. maxi-
serve several mining operations in the area. The balls produced in mum deviation from a straight line in 5 ft. Some rod users may
these early casting machines had many imperfections, such as large specify a tolerance of 49 in. in 5 ft. Straight rods are easier to charge
shrinkage and gas cavities, shifts or offsets at their parting line, and with a rod-charging machine; they are less prone to tangling of the
projecting sprues. However their production cost was quite low and charge, and are less likely to break prematurely in senice. They also
their performance in the relatively smalldiameter low-impact mills do a better job of grinding.
extensively used during that period was such that they were the most During the period when rod mills were first being tested as grinding
economical type of ball available. Due to steadily increasing produc- units, it was soon found that the rods should be made from a high-
tion costs, a demand for better quality. and a trend toward use of carbon steel, which had relatively good abrasion resistance and low
larger higher-impact mills, the use of white iron balls made on these ductility, so that when the rods were worn down to thin sections,
early types of casting machines has diminished rapidly since abovt they would break into smaller lengths without bending. When a tough
1950. or ductile steel was used in the rods, they would bend without breaking
In the late 1940s a new technique of casting, which would produce when worn thin, which in turn would cause serious entanglement
sound shrink-free balls, was developed by P. M. Payne.' The casting of the mam of rods in the mill. On the other hand, if one or more
techniques developed by Payne, together with certain modifications of the rods broke prematurely, this would also interfere with proper
and improvements, are now used for the production of substantial action of the rod charge and reduce the grinding rate and grinding
tonnages of balls down to the 1%-in. diam size. The development efficiency. Premature breakage usually occurred as a result of surface
of low-alloy cast steel balls, which also occurred in the late 1940s. or internal Aaws in the rod, though it could also occur as a result
was made economically feasible by use of the Payne casting machines. of excessive brittleness in the steel such as might be caused by a
Forged steel balls originally were used only in high-impact ball high phosphoruscontent. Consequently thequality control and inspec-
mills. which would break or spa11 white iron balls. The development tion requirements on grinding rods need to be fairly strict.
of automatic high-speed ball-forging machines, which occurred in For many years practically all grinding rods were made from
the early 1930s. substantially lowered the cart of producing forged- an unalloyed steel which contained 0.80-1.00% C, 0.304!.90% Mm,
steel balls and in some cases made them economically competitive 0.040% max P, and 0.050% max S. They were used in their as-
with white iron balls in low-impact mills. In the late 1930s. the abra- rolled condition with a hardness range of 240 to 300 Brinell. The
sion resistance and toughness of forged steel balls were further im- larger diameter rods tend to be softer than the smaller diameter rods.
proved by the use of alloying elements in the steel, coupled with a The AISI grade 1095 steel was usually specified. This type of steel
hardening heat treatment, which produced a fine-grained martensitic is still extensively used for grinding rods, though it is now being
structure in the balls. This improvement in quality further enhanced replaced at a number of operations by harder and more abrasion-
the competitive position of forged steel balls for grinding media. resistant grades of low-alloy steel.
Austenitic 12% manganese steel balls were used to some extent The first attempts to improve the abrasion resistance of grinding
in some of the early high-impact mills. The history covering the use rods involved the addition of small amounts of alloying elements.
of this steel for balls is obscure, but it appears that its use had been usually chromium, molybdenum, or vanadium, to increase the rod's
discontinued by the mid-1930s. The relatively high alloy content and as-rolled hardness. It is technically feasible to harden rods up to a
cost of this steel, together with its disappointing abrasion resistance, maximum of 380-400 Brinell by this p r o d u r e , but there are also
when compared to unalloyed or low-alloy high-carbon steel balls. some practical production problems involved, such as press-straight-
probably made the use of the austenitic 12% manganese steel unattrac- ening these higher-hardness rods prior to shipment. However there
tive. is currently some production of rods of this type, which reportedly
Martensitic alloyed white iron balls, principally of the Ni-Hard have from 15 to 25% better abrasion resistance than 1095 steel when
type, were developed during the mid-1930s. These martensitic white grinding high-silica ores.'
irons had much better abrasion resistance than the unalloyed or low- A more sucassful and popular approach to improving the abrasion
chromium (pearlitic) white iron balls. However, the use of these mar- resistance of grinding rods has involved the use of alloy additions,
tensitic white iron balls was impeded by the shortage of nickel and plus a special heat treatment in which the rods (as-rolled) are reheated
chromium during World War 11. After 1945, when alloying elements to a suitable heat-treating temperature (about 1500°F). then water-
again became readily available, them was a considerable expansion quenched in an arrangement which prevents the rods from warping
in the use of Ni-Hard for grinding balls, especially in the smaller while they are quenched. This produces a hard martensitic structure
sizes used for fine-grinding in cement mills. This has since been fol- through most of the volume of the rod, which is then stress-relieved
lowed by the commercial development and use of high-chromium and tempered to a hardness of 5 10 to 580 BHN. TO avoid premature
and chromium-molybdenum martensitic white iron balls for grinding breakage of these high-hardness rods in service, it is n-ry to
cement and some of the other son minerals. The development of maintain close control over the composition, surface quality and pro-
high-speed sand-molding machines has been an important factor in cessing of the steel. The heat-treated alloy-steel rods, as supplied by
the economic production of the smaller sizes of these alloyed white US manufacturers, cost I2 to 15% more than the 1095 steel rods,
iron balls. (1974 prices). As is indicated in Table I, the 1095 steel rods wear
For primary grinding in ball mills, there has been a trend toward at rates which are from 36 to 191% faster than the alloy rods.= It
decreasing the diameter of the balls. This is due to various factors, appears, therefore, that the use of the alloy rods can be economically
such as finer products from the crushing plant which produce feed justified in rod-milling operations.
for the grinding mills, the use of rod mills for first-stage grinding, Matenah for Grinding Balls. Grinding balls for tumbling mills
and the recognition that the best grinding efficiency is obtained when range in size from 5-in. maximum diameter down to about %-in.
the balls are no larger than that needed to effect initial breakage of minimum. Currently there is very little production of balls larger
the largest particles of rock or mineral fed to the mill. Up to about than 4-in. diam, except for semiautogenous mills. Balls smaller than
1940, there was a substantial demand for C and 5-in. diam balls, %-in. diam may be supplied on special order from the larger suppliers.
3C-6 GRINDING
Table 1. Consumption of Grinding Rods in Various Ores scale tests in the relatively high-impact ball mills at Climax and later
in the USA' at Morenci, where they were found to provide an excellent combina-
Wear rates, tion of abrasion resistance, toughness, and economy. Subsequently
Ib per ton several other steel ball producers have installed the equipment needed
Type of Rod Difference, to produce the heat-treated, highcarbon, low-alloy martensitic steel
ore diam. in. Alloy High carbon %
balls, which now account for the major tonnage of balls produced
Taconite 4 0.82 1.25 52 for the mining and cement industries in North America. Most of
Ilmenite 2% 0.43 1.25 191 the forged and cast-steel balls presently produced in the United States
Lead-zinc 3':s 0.83 1.15 39 have this highcarbon martensitic composition and structure.
Copper 4 0.65 0.90 38 In the late 1930s techniques were developed for running wear
Uranium 2'/r 0.58 0.90 55
tests on grinding balls whereby a number of groups of balls, each
Copper 4 0.61 0.83 36
representing a variation in composition, structure, or source, were
Copper 3 0.36 0.53 41
suitably identified and then run all at one time in a ball mill to provide
' Data from Armco Steel Cop. a direct comparison of their abrasion and impact resistance. 0. W.
Ellis' and his associates at Ontario Research Foundation pioneered
However, some of the smaller firms produce grinding balls ranging this technique, where the tests were run in laboratory-size mills. His
in size down to %-in diam. Recently, small cylindrical metal shapes, tests were followed by those of T. K. Prentice; and T. E. Norman.6
about %-in. diam, called Minipebs have been developed for ultra- who conducted their tests in commercial-size ball mills under normal
fine grinding of cement. Similar small sizes and shapes made from conditions of ball mill operation. The techniques used for these tests
high-alumina ceramics are also available. have since been adopted by other investigators, in some cases with
The size of the balls and the amount of repeated impact to which the addition of radioactive tracers for ball identification. The results
they are exposed during their service in a mill are important consider- from these various studies have been very helpful in defining the
ations in the selection of suitable materials. Cost is also a very impor- factors and studying the variables which influence ball w a r . They
tant consideration. During the early development of tumbling mills, also evaluate resistance to breakage or spaUing of the balls in actual
the choice of suitable materials for grinding media was quite limited. service conditions and can be very helpful in determining the most
For low-impact mills the usual choice was either white iron balls economical type of ball for a particular milling operation.
or flint pebbles. For high impact mills, which normally used the larger- An example of the type of information obtained from these wear
diameter balls, the usual choice was either high-carbon forged steel testing techniques is provided in Table 2 from a paper by Norman
used in its as-forged condition or austenitic manganese steel which and Hall.' In this table the low-alloy highcarbon heat-treated mar-
could be made into balls by either casting or forging, then heat treated. tensitic steel balls were used as a comparative standard with an as-
There is also some record of the use of small amounts of chromium signed relative wear rate of 100. The other materials listed were
in the high-carbon steel balls. Gray cast-iron balls may also have compared to this standard. On the basis of present (1974) cost-wear
been used, though there is little record of this. Balls and other shapes ratios, the standard forged steel (item 2) and its companion cast steel
made from steel scrap such as worn railroad rails, broken drill steel, (item 3) are the most economical for use in the Climax mills. However
and other cheap sources were also used to some extent. the good performance of the LSCr-3Mo white iron (item 7) at Climax
Alloying of white iron balls with small amounts of chromium led to the investigation of this type of iron for possible use in other
and the practice of casting in metal molds to improve their abrasion conditions of operation. It was found that this type of iron, which
resistance was fairly well established by the mid-1930s. At about has a structure of hard and relatively tough chromium carbides in
this time, Ni-Hard, which is a nickelchromium white iron with a a martensitic matrix, had outstanding abrasion resistance when grind-
hard martensitic matrix, was developed and used to a limited extent ing minerals softer than quartz, which in turn made it economically
for grinding balls. Later, after World War 11, there was considerable feasible for commercial use, in spite of its higher first cost. When
expansion in the use of Ni-Hard for grinding balls. used for smalldiameter balls, it is permissible to reduce the molybde-
An important step fornard in the development of high-quality num content in the composition down to 1% or I s s , which effects
steel balls occurred in 1937 when Sheffield Steel Corp., in Kansas some saving in cost. Commercial production of balls made from this
City, installed processing equipment to produce heat-treated martens- type of iron began in about 1960, principally in Europe, where it
itic low-alloy forged-steel balls. Balls of this type were run on large- was found that they had from three to seven times the abrasion resis-
Toble 2. Relative Wear Rates of bin. Grindine Balls in Primorv Ball Mills ot Climax
Relative wear rate
Typical composition, %
Item Hardness,' Marked Large
No. Descriotion and heat treatment C Mn Cr Mo Ni RC balls scale
For& or b s t St--I
1. High Cr, tool steel, OQ&W 2.15 0.3 12.0 0.7 ... 61 92 ...
2. Low-alloy forged, WQ&TP 0.80 0.6 0.3 0.2 ... 64 100 100
3. Low-alloy, east, WQ&T 0.71 0.6 0.4 ... ... 62 102 ...
4. CrMo cast, a s a s t , Agl 1.20 0.7 1.2 0.2 ... 41 114 122
5. Unalloyed forged, OQ§ 0.75 0.7 ... ... ... 40 117 119
6. Lowearbon forged, air cool 0.20 0.5 ... ... ... . . .n 204 ...
Cmst Whlt. Irons
7. Cr-Mo, 1 Malloy AQ&Tt 3.00 0.8 15.0 3.0 . .. 68 82 ...
8. Ni-Cr iron. sand cast 3.25
9. Low Cr iron. sand cast 3.00
Table 3. Connorciol Type. of Grinding Balls P r o d u d In the USA and Canada (1974)
Typical composition, 96'
Type Hardness,
No. Description C Mn Cr Mo Ni Rc
1 High-chromium iron (cast) 2.44.4 0.61.0 12.&!20.0 0.1-1.0 - 62-68
2 Ni-Hard (east) 2.M.5 0.3-0.6 1.1-1.5 - 2.84.0 55-60
3 Martensitic low-alloy steel 0.61.0 0.4-1.0 0.3-0.8 0.0-0.2 - 61-64
(forged or cast)
4 Low-alloy iron (cast) 3.0-3.5 t t t t t
5 Pearlitic higharbon steel 0.6-0.9 0.6-0.9 - - - 3642
(forged)
6 Pearlitic white iron (east) 2.83.4 0.3-1.0 0.5-2.0 0.0-0.5 O.bl.0 4&50
Actual compo8ition range from an ind~v~dual producer will be narrower than the ranges shown.
r T.v.~ 4e represents proprietary and exper~mentaleompo8itions made by several pmducers. Composit~on,
structure and ha;dness va2es acmrdingly.
Note: Hardness will vary with hall diameter, the larger diameter balls being softer than the smaller
diameter halls.
tance of forged steel balls when used in cement grinding. Production rate than type 3. It has insufficient impact resistance for use in the
in the United States was started about six years later. At present modem, largediameter grinding mills. Also it is no longer a cheap
balls of this type are used quite extensively at cement plants in Europe, material. For example, up to about 1950 the cost per pound of white
the United States, and some other countries. iron balls was less than half that of forged steel balls. At present
Literally hundreds of variations in composition, structure, heat the white iron balls, if and when available, cost almost as much per
treatment, and production methods of grinding balls have been investi- pound as forged or cast steel balls. Currently, the principal production
gated over the past 60 years. Quite a few of these have been tested of white iron balls in the United S t a t e and Canada is by two large
or used commercially. However there are only a few surviving types mining companies, who produce the balls in their own foundries for
being produced at the present time in the United States and Canada. use in their own grinding operations.
Table 3 lists the six general types together with their normal wmposi- While no published statistics are available, it appears that the
tion and hardness range. They are listed in their approximate order current demand for grinding balls in the United States and Canada
of decreasing abrasion resistance, though minor changes in this order exceeds 300,000 tpy. They are priced and sold in a highly competitive
may occur, depending on the characteristics of the ore or mineral market. In spite of this, the cost of steel balls in the popular 2 to
being ground and operating conditions in the mill. 3-in. sizes has risen from a low of about 3# per Ib in the mid-1930s
The high-chromium iron, type 1 in Table 3, is presently used up to a range of 10 to 156 per Ib in 1973-74. Obviously the need
principally in dry grinding applications in the cement industry in for high abrasion resistance in these balls is much greater at present
the United States and Canada. It has displaced most of the Ni-Hard than it was in the mid-1930s.
balls (type 2) and some of the low-alloy steel balls (type 3). formerly Table 4 lists the approximate current relative prices of balls in
used in these cement-grinding applications. the three most popular grades and the various s i m produced.' These
The martensitic low-alloy steel composition, type 3, which includes relative prices are dependent on the cost of production, which tends
both the forged and cast grades, constitutes the principal tonnage to increase for the forged balls as the size decreases or increases
of balls used in the United States and Canada at present. Due to from the popular 3-in. size. Also the cost is influenced by the alloy
its favorable cost-wear ratio, combined with its adaptability to most content of the balls, which is normally less in the small-diameter
milling conditions, the demand for this type of ball has increased balls than in the largediameter balls.
steadily since its introduction in 1937. The 0.6 to 1.0% carbon range Nonferrous Grinding Media. During the early development of
listed for these type 3 balls represents a range from the various availa- grinding mills, various types of natural stone pebbles were used or
ble sources and various sizes of balls. Carbon contents near the high tested for grinding media. It was found that most types of rock tended
side of this range are desirable for best abrasion resistance. However. to break or spa11 in these mills, so that consumption of this type of
high carbon contents can also reduce the toughness of the balls and grinding media was excessive. However it was also found that flint,
make them more susceptible to spalling or breakage, so a compromise which is a fine-grained quartzite, had good durability and wuld be
may be nacessary. In recent years, several producers of these balls used economically as grinding media. The k t quality flint pebbles
have found that by close control over steel quality and heat-treatment, were found on the beaches in Denmark and France, so these were
the high side of the carbon range can be used quite safely for most imported and used quite extensively for fine grinding in low-impact
of their production. Eventually the carbon content of commercially mills up to about the beginning of World War 11. There is still some
produced balls in the type 3 classification may be increased above use of these pebbles for grinding media, but increases in their cost
1% to obtain a further improvement in abrasion resistance. and limited availability have greatly reduced their use in North Amer-
The low-alloy cast iron, type 4, represent several new compositions ica. As a supplement, there is some use of flint pebbles and quarried
which have been under development by several producers of cast fine-grained quartzite obtained from domestic sources.
balls during the past few years. While there is presently sufficient Some ores contain sufficient flint or similar fine-grained rock to
production of these types to merit their listing in Table 3, they still serve as fine-grinding media for that particular ore. The gold ores
constitute a relatively small percentage of the total ball market. At on the Rand in South Africa are of this type and are used in this
least one of the compositions included in type 4 has sufficiently good manner. This practice approaches autogenous grinding, which will
impact resistance to make it competitive with steel balls in high-impact not be considered here.
mills. Porcelain balls were also found to be suitable for fine grinding,
The pearlitic high-carbon steel, type 5, was the principal material especially in mills where iron contamination by wear or corrosion
used for forged steel balls up to the time the martensitic low-alloy of iron or steel balls wuld not be tolerated. About 20 years ago, a
steel balls were developed. During recent years the martensitic grade high-alumina ceramic became available at prices which made balls
has largely displaced this pearlitic steel, which normally wears about from this material economically competitive with procelain balls and
20 to 35% faster than the martensitic grade in ore-grinding operations. natural pebbles. High-alumina ceramic grinding media normally wear
The pearlitic white iron, type 6, has now lost most of its early at a much slower rate than either porcelain balls or flint pebbles
popularity, due to the fact that it has a substantially faster wear and have the advantage of a higher specific gravity, which in tum
Table 4. Approximate Relative Prices of Various Sixes and Grades of Grind-
ing Balls*
Relative price per unit weight
Diam, in. Low-alloy steel Carbon steel Highshmmium iron
5/8 234
228
Ye 142 224
1 118 218
1% 101 216
1 'i2 106 95 214
2 117 106 212
2 'h 106 103 236
3 100' 87 234
3% 140 110 232
4 117 106 230
Prices are relative to an assigned value of 100 for the Sin. alloy steel balls. Relative
prices am subject to change depending on producer and market conditions.
produces a better grinding rate than would be obtained from flint a 3-in. ball makeup and others used a 50% 3-in. and 50% 2-in.
pebbles or porcelain balls. ball makeup, confirmed this observation.
About 1950, sintered tungsten carbide grinding balls were devel- Certain factors can cause a departure from the rule that the wear
oped. While their first cost is relatively high, their combination of of individual balls in a charge varies as DZ.For example when new
very high abrasion resistance, high specific gravity, and freedom from steel balls are added to a mill, they normally have a decarburized
contamination of the ground product makes these balls attractive surface which may be from 0.010 to 0.060 in. thick. While this decar-
for use in small mills grinding special products, which include tungsten burized shell is being worn off the ball, the ball wears at a faster
carbide powders. rate than it does after the decarburized surface has been removed.
Table 5 provides some comparisons between the ferrous and non- Also if the largest balls in a mill are segregated towards the feed
ferrous grinding media which are commercially available at present. end of a mill, by use of segregating linings or a conical-shaped mill,
The relative costs in this table are based on 1974 prices FOB the then these large balls are abraded faster by coarse particles in the
producer's plant, except in the case of the imported pebbls, which feed. Other conditions of mill operation may also cause size segrega-
are based FOB the port of Houston, Texas. tion of balls in the tumbling charge, which in turn may cause their
Wear Rates of Grinding Balls. By direct measurement of the wear rate to depart from the Darule. However for practical purposes
wear of grinding balls in commercially operating mills, several in calculating the size distribution of balls in a charge, it may be
invstigators5. have obtained conclusive evidence that the individ- assumed that the balls l o x diameter at a constant rate, so that in a
ual balls in any specific ball charge tend to wear (lose weight) in seasoned charge in which a certain number of balls of one diameter
direct proportion to their surface area. This means that the balls was added at regular intervals (i.e., daily) there will be an equal
lose diameter at a constant rate throughout their useful service life. number of balls of each size interval. In most grinding mills, the
It also means that individual ball wear varies as DZ,where D is balls change from their original spherical shape to concave-faced poly-
the diameter of the ball at the time its wear rate is measured. This gons, disks, and other irregular shapes after they have lost about
concept of ball wear has been disputed by others who reasoned that two-thirds of their original diameter. If an ex- of these small irregu-
balls should wear in proportion to their weight or as DS, since, for lar shapes builds up in the ball charge, its grinding capacity will be
example, a charge of 50 tons of 3-in. balls in a mill would wear at reduced, sometimes by as much as 10. However in wet grinding mills
practically the same rate as a 50-ton charge of 2-in. balls, even though these small shapes tend to float out with the mill discharge, or in
the 2-in. ball charge had a larger total surface area. Normane explained grate-discharge mills they pass out through the openings in the dis-
this latter observation by pointing out that when the total surface charge grates. In cement mills, which have a relatively large length-
area of the ball charge was increased by the change from 3- to 2- to-diameter ratio, these small shapes, which are called h s b or spit-
in. balls, the wear per unit of surface area of each ball in the charge zers, tend to accumulate in the mill. Periodically, when a noticeable
dropped in direct proportion to the increase in total surface area of drop in grinding capacity has occurred, the mill charge is dumped
the charge. His measurements of wear rates of individual balls in and the "trash" is screened out.
the 9 x 9-ft primary mills at Climax, where some of the mills used A seasoned charge of steel balls, containing equal numbers of
Table 6. Determination of Initial Ball Charge to A New Mill* their service life. This varies over a wide range, depending on the
abrasion resistance of the balls, the abrasiveness of the material being
(a) Assume that 100% 3-in. balk will be charged to a mill on a periodic ground, the size distribution, and total surface area of the ball charge,
basis and mill holds 6Cton balls. Method to calculate initial ball charge
for new mill follows (ball diameters commercially available from steel and the diameter of the mill. The effect of mill d i e t e r is not generally
manufacturer used): appreciated. This effect is due to the fact that the weight and surface
area of a ball charge in a mill increases as the square of the mill
Diameter, diameter. The energy input to tumble this charge increases as the
in. Thm. wt % Wt. tons
2.6 power of the mill diameter, so energy input per unit of total
ball area of the charge increases as the 0.6 power of the mill diameter.
Since there is direct relationship between energy input and ball wear,
it follows that ball wear increases as the 0.6 power of the mill diameter.
This statement assumes that the ball charge in each mill occupies a
given percentage, such as 50%. of the mill volume. It should also
be recognized that even though individual balls lose diameter faster
as mill diameter increases, the wear, when expressed in pounds per
ton of ore ground or pounds per kilowatt-hour, does not n e s a w i l y
(b) Assume multiplesize makeup addition for initial charge determination. increase.
Assume that 709-30%, 2 and 3 in., respectively, will be recharged on
periodic basis. By use of direct-measurement techniques on suitably identified
balls, Norman1, had determined the wear rate and life of balls in
Calculation follows: 70 Ratio of 2 to 3-in. 4.00 wt of 3-in Ball - a number of mills operated under widely diverse conditions. His results
36 wt of 2-in Ball -
are listed in Table 7. The wear rates are for 0.80% carbon, low-
There are 7.84 2-in balls to one 3-in ball alloy martensitic steel balls, fully hardened through at least 95% of
Theo. Ball wts their original volume. The data in Table 7 indicate that the life of
Diameter, Balanced the balls, 2- to 3-in. original diameter, in the commercial-size o r e
in. 3-in. 2-in. wt % Wt, tons grinding mills, ranges from a minimum of 73 days to a maximum
of 312 days. In a 3-ft diam test mill grinding feldspar and calcite,
the lives on 3-in. balls were 608 and 7,260 days, respectively. In
dry-grinding cement clinker, a calculated life of 13,500 days was ob-
tained on 4-in. balls in the primary compartment and, at a different
plant, the calculated life on 1U-in. balls in the second compartment
was 2.000 days. These life figures on balls are of particular interest
in estimating the time required to completely displace an old charge
of balls when a test is started on balls made from a new composition
This material is based on theory that ball wear is proportional to diame or new source.
tor squared, which is to say that regardless of the number of sizes
of balls in the charge, each piece will lose an equal thickness of material Another way of determining the time required to displace a charge
from its surface in a given period of time. of balls in a mill has been proposed by A. L. Bard." This is illustrated
in Table 8. In the calculation accompanying Tal~!e 8, it is indicated
' Dependent on ball makeup size to the mill. that where x tons of balls are maintained in a mill charge by periodic
After A. Souder, Armco Steel Cop. (daily) additions, then a total of 2.9 times x tons will have to be
balls in each size interval will weigh about 300 pcf. By increasing added to the mill by the periodic makeup before the original charge
the proportions of smaller-size balls in the charge, densities up to in the mill has been completely displaced. Bard's calculations are
about 325 pd are attainable. In a charge where all of the balls are based on the assumption that the balls in the mill lase diameter at
the same diameter, the density is 282 pcf for steel. a constant rate. His calculations agree well with the data from actual
Some mills will grind more efficiently to the desired product size field tests and with the ball-life estimata in Table 7. Bard has made
when a higher proportion of small balls exist in a seasoned charge a similar calculation for a charge of grinding rods, as illustrated in
than would naturally be obtained when the periodic makeup balls Table 9. Here the makeup factor is 2.07 times x.
are ail of one diameter. This was conclusively demonstrated by Barker Corrosion Influence on Grinding Media h u m p t i o n . There is
and Lewis,lo in the ball mills at Ajo, Ariz. They found that a certain considerable disagreement and probably some misunderstanding over
percentage of the larger diameter balls was needed for initial break- the role that corrosion plays in the wear of grinding media. It is
down of the larger or tougher pieces of mill feed, while a higher well known that when the raw feed to cement plants is ground wet,
percentage of smallerdiameter balls in the charge produced the most the wear of grinding media and liners is usually from five to ten
desirable screen analysis of the ground product. This condition exists times the wear rate when the feed is ground dry. Also in cement
at many single-stage grinding operations. To obtain the desired screen finishing mills, which always grind dry, the wear rate, when expressed
analysis of the mill discharge where these conditions exist, size ration- in weight loss per day or weight loss per kilowatt-hour, is very low
ing of the periodic ball makeup may be used. For example, one large when compared to wet grinding operations. These observations are
singlestage milling operation on a -%-in. copper ore uses a daily well confirmed by the studies of Wesner et al.,13 Bond,&' and other
ball makeup ration consisting of about 10% 341-in., 40% 3-in., and investigators. It is logical to assume that in wet grinding, a freshly
50% 2-in. balls (weight percentages). This ration was found to be abraded metal surface reacts with the water to form an oxide or
the most desirable after several years of testing with various propor- hydroxide surface coating and hydrogen gas. This freshly formed
tions of the three sizs. A discussion of the many considerations in- coating is not usually very adherent or abrasion-resistant, so it is
volved in ball size rationing is given by Crow." easily abraded off the surface of the ball, rod, or liner, thus permitting
When a new mill is charged with balls, it is desirable to proportion further chemical reaction between the water and the metal surface.
the weights of each size charged to the proportions which will exist There are indications that chromium oxide or hydroxide films are
in the mill after it has operated for a sufficient length of time to more abrasion-resistant than iron oxide or hydroxide films, since high-
produce an equilibrium or seasoned charge. Table 6 illustrates how chromium martensitic steels and irons tend to resist abrasive wear
such initial charges can be calculated when the daily ration to the in both wet and dry grinding substantially better than low-chromium
mill after it is put into operation will be all 3-in. balls, and, when manensitic steels. However even high chromium contents fail to pro-
the daily ration will be 70% 2-in. and 30% 3-in. balls.3 tect the softer grades of stainless steels from abrasion in grinding
The rate at which balls lose diameter by wear in a mill determines mills.
GRINDING
Table 7. Wear Rates of 0.8 Percent Carbon Martensitic Steel Balls in Various Mills
Mill Ball Diameter Approx.
Plant diam, makeup loss per life,t
no. Abrasives' ft diam, in. day, in. days
la Crushed Mo ore-high quartz 9 100% 3 0.0228 99
lb Crushed Mo orehigh quartz 9 50% 3 0.0172 131 on 3 in.
50% 2 0.0172 73 on 2 in.
lc Crushed Mo orehigh quartz 6 100% 3 0.0134 168
2 Rod-milled gold ore-hard silicate 5 100% 2 0.0090 139
3 Crushed Cu ore-feldspar & quartz 6% 100% 3 0.0086 262
4 Crushed Cu orequartz & feldspar 8 100% 3 0.0105 214
5a Pea gravelsandquartz & feldspar 3 All 3 charge 0.0083 2-71
5b Crushed feldspar 3 All 3 charge 0.0037 608
5c Crushed calcite 3 All 3 charge 0.0003 7260
6a Crushed Cu orequartz & feldspar 6% 100% 2 0.0110 114
6b Rod-milled Cu ore-quartz & feldspar 12% 100% 2 0.0089 140
7 Crushed Cu orequartz & feldspar 10 50% 3 0.0126 179 on 3 in.
50% 2 0.0126 99 on 2 in.
8 Rod-milled Mo orequartz & feldspar 12!$ 100% 2 0.0087 144
9 Rod-milled Cu ore-quartz & feldspar 7 100% 2 0.0115 109
10 Crushed Fe orehematite 5 100% 2 0.0040 312
11 Rod-milled Fe ore-magnetite & quartz 11% 100% 2 0.0074 169
12 Rod-milled Fe oremagnetite & quartz 14 100% 2 0.0040 312
13 Cement clinker, dry, primary compart. 8 4 & smaller 0.0002 13,500 on 4 in.
14 Cement clinker, dry, second compart. 13 l l h & smaller 0.0005 2.000 on 11/2 in.
' All wet grinding in water slurry except cement clinker.
t Life is calculated as the time to wear balls from their original diameter down to 3A-in. diam, except on lX%-in.
balls, which were assumed to be rejected a t K-in. diam.
In wet-grinding o f most ores, where wear rates per unit of surface cles continue to react with water, probably until they are completely
area are relatively high, substantial reductions in wear rates can be converted to a n oxide o r hydroxide.
obtained by increasing the h a r d n e s of the metal in the grinding media A clue to the proportions of wear which occur by corrosion and
and by the introduction of hard carbides (FeaC o r CrrCs) into the by abrasion, respectively, may be found in the wear r a t e , in terms
structure of the metal. This indicates that a t least part of the wear of diameter-loss per day, which are listed in Table 7. Here the diameter
occurs by gouging o r scratching of the metal underneath the oxide loss of the balls ranges from a high of 22.8 mils per day d o w n t o 4
surface-films. T h e fact that metallic particles a r e actually removed mils per day in wet ore grinding in commercial size mills. T h e balls
from the surface is also indicated by measurement of hydrogen evolved were all grinding in neutral o r slightly basic pulps s o their corrosion-
from freshly ground pulps. Hydrogen evolution has been observed
Table 9. Calculation of Weight of Periodic Rod
for several hours and sometimes for a day o r two after the pulp
Additions Required to Displace the Original Chorge
has been ground, which indicates that the freshly abraded iron parti-
in a Rod Mill
Rod Rod wt, Cum. w t loss, Cum.
Table 8. Calculation of Weight of Periodic Ball No. diam, Ib, per % of Ib per % of
Addiions Required to Displace the Odginal Chage in Days in. ft total rod total
a Bail Mill
Start
Units of Ball % Wt %
time d i m , in. Wt, g cum. loss, g cum.
Start 2.0 537
1 1.9 461 7 6 14.2
2 1.8 392 47.0 69 27.0
3 1.7 330 58.1 62 38.5
4 1.6 2-75 67.4 55 48.8
5 1.5 22-7 75.1 48 57.7
6 1.4 184 81.3 43 65.7
7 1.3 148 86.3 36 72.4
8 1.2 116 90.2 32 78.4
9 1.1 89 93.2 2-7 83.4
10 1.0 67 95.5 22 87.5
11 0.9 49 97.1 18 90.9
12 0.8 35 98.3 14 93.5
13 0.7 23 99.1 12 95.7
14 0.6 15 99.6 8 97.2
15 0.5 8 99.9 7 98.5
16 0.4 4 100.0 4 99.3 20 1 2.67 100.0 0.71 91.9
Purge 4 100.0 Purge 0.67 100.0
rn m m m
Courtesy Armco Steel Corp. Courtesy Armeo Steel Corp.
Wt of ball charge, 2,960 g; l i e of 2-in. ball 16 units of time; Wt of md charge. 316.03 Ib; life of 3%-in. md, 20 days; wt
w t loss per unit of time 537 g; wt 16 pieces 2-in. balls: 16 loss per day, 32.71 Ib. Wt of 20 3%-in. x 1-ft rods = 654.2
x 537 = 8,592 g. 8,59212,960 = 2.9. Therefore, the indication Ib. 654.21316.03 = 2.07. Therefore the indication is that
is that 2.9 times the weight of ball charge would have to 2.07 times the weight of md charge would have to be
be added, as makeup, to displace charge. added, a s makeup, to displace charge.
TUMBLING MILL CHARACTERISTICS
erosion rate should remain relatively constant. Actual rate of wear data from 5 large taconite ore operations in the United States. These
by corrosion-erosion is not known, but it would appear to be less operations all used two-stage grinding circuits in which one or two
than 4 mils per day and might reasonably be of the order of I mil ball mills ground the product from each rod mill. In most cases
per day. Under these circumstances. 25% of the wear would be due the rod mills had used high-carbon steel rods for many years then
to corrosion-erosion at plants 10 and 12, while only about 4% of switched to the use or testing of the alloyed martensitic steel rods,
the wear would be by corrosion-erosion at plant la. so consumption data on both t y p s are listed separately.
The very low wear rate of 0.3 mils per day in the 3-ft-diam mill Table I I lists the reported consumption of grinding balls in I2
at plant 5c when wet-grinding calcite, is interesting. In this test the different types of milling operations involving a total of 147 milling
ground pulp was definitely alkaline and the balls developed a high plants. Data on the crushed and the rod-milled nonferrous ores are
polish. It appears that wear due to corrosion-erosion was also very mainly from the large copper and molybdenum ore-milling operations
low. Similar low wear rates were obtained in the dry grinding of in the United States and Canada which all used the type 3 (martensitic
cement clinker at plants 13 and 14 (Table 7). Under these circum- steel) grinding balls. For taconite ore grinding, which all used rod-
stances, it is difficult to determine whether the wear was due to corro- mill grinding of magnetic taconites ahead of the ball mills, about
sion, erosion, or abrasion. Theoretically the relatively soft minerals one-third of the weight of the rod-mill discharge, representing about
in cement clinker are not capable of scratching fully hardened martens- one-half the silica in the ore, was removed by magnetic separation
itic steel, so possibly much of the wear in clinker grinding is due to prior to ball milling. This removal of silica substantially reduces the
corrosion-erosion even though the corrosion rate in dry grinding is rate of ball wear in the taconite mills. Also, the reported rates on
probably much less than in wet grinding. line 3 in Table I I are per ton of rod mill feed. If they had been
The influence of corrosion-erosion should be considered when wear reported per ton of actual feed to the ball mills, the ball consumption
tests on ferrous alloy balls are run in small-diameter mills. For exam- per ton would be about 33% higher. Only type 3 balls were used
ple, by using the rule that the specific wear rate (diameter loss) on in these taconite mills.
individual balls in a normal mill charge varies at the 0.6 power of Lines 4 and 5 in Table I I illustrate the marked reduction in wear
the mill diameter, it follows that the balls lose diameter about four rates in raw cement-rock grinding when a change is made from wet
times faster in a 10-ft diam mill than in a I-ft diam mill. The percentage to dry grinding. The wear in wet grinding averages from seven to
of the loss due to corrosion-erosion is therefore much greater in the nine times faster than in dry grinding. Lines 4 and 5 include the
I-ft mill. Consequently, when the relative wear resistance of several wear rates from five different t y p s of grinding balls, which probably
ball types or compositions is determined in a small laboratory mill, explains why there is a very wide spread between the highest and
the results may not correlate well with the relative wear rates obtained lowest rates. The abrasiveness of the raw feed to cement plants also
in larger commercial-size mills grinding the same ore or mineral. It varies widely, depending on local conditions.
is recommended that where relative wear rates of different t y p s of Lines 6 and 7 in Table I I, which cover the finish grinding of
balls are to be determined, the balls be tested in the full-size operating cement, also show a wide spread between the highest and lowest
mills where their commercial use is contemplated. The marked-ball wear rates. In general, the mills using the highchromium iron balls
technique of testing developed by Norman6 or the radioactive-tracer (type I in Table 3) provided the lowest wear rates. It is also obvious
technique described by Wesner et al.12 can be used for such tests. from the pound per kilowatt-hour values, that the abrasiveness of
Consumption of Grinding Media in Rod and Ball Mills. The dry cement clinker is much less than that of the wet ores listed on
consumption of grinding rods in I1 large operations milling low-grade lines 1, 2, and 3.
copper, molybdenum, lead-zinc, gold, uranium, and ilmenite ores in Bond" has made extensive studies of the abrasiveness of various
North America is listed in Table 10, together with corresponding ores and minerals. By determining the work index ( Wi) and abrasion
index ( A i ) in a specially designed laboratory machine, he is able to operation.' This consumption rate is obtained with low-lift liners.
predict the wear of rods and balls in specific grinding operations. It was found that when lifter height on liners was increased, pebble
His equations on rod and ball wear are as follows: consumption increased to about 20 Ib per ton. The increase was due
Wet rod mills: principally to increased breakage and spalling of the pebbles.
Rod wear in lb/kw-hr = 0.35 (Ai-0.020)0m Tests have also been conducted at Climax on jasperite stone peb-
bles obtained from Minnesota. These were tested in competition with
Wet ball mills: flint pebbles imported from France. The jasperite stone showed a
Ball wear in Ib/kw-hr = 0.35 (Ai-0.015)1'3 slightly lower consumption rate than the flint pebbles, but the lower
consumption rate was insufficient to justify the higher delivered cost.
Dry ball mills: A similar competitive test was run between high-alumina ceramic
Ball wear in lb/kw-hr = 0.05 fi balls and flint pebbles in the Climax mills. The ceramic balls improved
the grinding rate in the mills and were consumed at about half the
The foregoing equations were developed from the actual wear rate of the flint pebbles. However, these improvements were insuffi-
data supplied from a large number of grinding operations. Since there cient to justify the extra cost of the ceramic balls at Climax. It should
has been a continuing trend toward the use of more abrasion-resistant be noted, however, that these mills could not beoperated under condi-
rods and balls since the collection of Bond's data, it will be necessary tions of high pulp density and slow speed, as recommended by the
to reduce the constants in Bond's equations to bring them in line manufacturer of the ceramic balls. Tests at other fine-grinding opera-
with this trend. Work along these lines has also been conducted at tions indicate that where favorable operating conditions can be main-
the grinding research laboratory of CF&I Steel Corp. in Pueblo, Colo. tained, the ceramic balls may have up to ten times the abrasion resis-
The relative wear rates, as developed by Bond, are listed in Table tance of flint pebbles.
12. Where comparisons exist, it may be seen that Bond's wear rates. Selection of Grinding Media. In selecting materials for grinding
in pounds per kilowatt-hour, are slightly higher than the correspond- media, it is obviously n-ry to select a material which will not
ing wear rates from more recent data listed in Tables 10 and 11. break or spa11 excessively in the mi11 where they are to be used.
This is probably due, as previously mentioned, to the continuing trend When this requirement has been met, the next major consideration
toward use of more abrasion-resistant materials in the rods and balls. is usually the cost-wear ratio. A third consideration is the density
Consumptionof Grinding Medh in Pebble M i . Currently p e b of the media, since the grinding rate and power consumption of a
ble mills, other than autogenous mills, are used principally for regrind- mill tends to vary in almost direct proportion to the density of the
ing concentrates or for fine grinding where contamination by corrosion grinding media.
or wear of metal balls must be avoided. In the pebble mills operated In selecting media which will not break or spall excessively in
prior to World War 11, which were used mostly for fine-grinding service, the type 5 pearlitic steel in Table 3 may be considered as
gold ores. Taggart15 indicates the consumption rate on good-quality having the highest toughness. However its relatively poor cost-wear
Danish or flint pebbles was in a range of 0.5 to 8.0 Ib per ton of ratio makes it an unattractive choice for most milling operations.
ore ground. Average consumption was about 2 Ib per ton. Most of The type 3 martensitic steels follow next in order of toughness, pro-
the pebble mills were about 5 R diam and 16-22 ft long. Impact vided they are made with a fine grain size under properly controlled
conditions were low, which favored the use of pebbles. When hard conditions of manufacture. These proper eonditiom of manufacture
ore was used as the grinding media in similar mills, its consumption have been adopted and rigidly enforced by all of the major producers
rate was about 50 to 80 Ib per ton. In South Africa, the tube mills of grinding balls in the United States so that the type 3 steel can
on the Rand consumed about 200 to 300 Ib of lump ore per ton of now be safely used in practically all of the existing ball-milling opera-
ground product. Frequently combinations of lump ore and iron or tions. Use of this type of steel in grinding rods posed a more difficult
steel balls were used together in these tube mills. problem for the producers, but indications are that the initial produc-
In the United States there has been a substantial use of flint pebbles tion problems, which frequently lead to breakage of rcds made from
for regrinding rougher concentrates, in which quartz is the principal this type 3 steel, have been solved. It now appears that the heat-
abrasive, at the operations of Climax Molybdenum Co. The mills treated martensitic steel rods can be safely used in most rod-milling
are 61,4 and 8 ft diam by 20 ft long, equipped with Iow-discharge operations. The advantage of the martensitic steel balls and rcds lies
grates. The impact conditions are relatively high in these mills, so in their attractive cast-wear ratios.
consumption of flint pebbles, 2 to 2% in. diam, is correspondingly The martensitic steel balls are currently made both as forgings
high, averaging about 9 to 12 Ib per ton of f d to the regrinding and castings. The composition, heat-treatment, and quality control
Table 12. Average Abrasion Index of Materials and Wear Rates Produced in Rod
and Ball MIIIs*
Weat rate, Ib per. kw-hr
No. Material Wet rods Wet balls Dry balls
-
procedures on each are similar. The forged balls are more free from tant source of grinding media. Flint pebbles are normally obtained
physical defects such as surface appearance, while the cast steel balls from the beaches in Denmark and France or from a new source
are normally more free from surface decarburization. The surface which has been developed more recently in Texas. Jasperite stone,
decarburization on forged steel balls can reduce their overall abrasion which is quamed in Minnesota, is a finegrained quartz similar to
resistance by several percent, especially on the smallerdiameter balls. flint. The low specific gravity of flint, when compared to steel, together
However where largescale tests have been conducted to compare with its susceptibility to breakage and spalling in high-impact mills,
the wnsumption rate of the forged balls to cast balls, there appears have effectively ruled out its use for grinding media in the modem
to be very little, if any, difference between the two types.' In high- largediameter, high-capacity primary grinding mills. However in view
impact mills, the forged balls appear to be somewhat less susceptible of its low cost per pound relative to steel, it is still a potential competi-
to breakage or spalling than the cast balls. The forged steel balls tor to the martensitic steel or white iron balls for the regrinding of
currently are made regularly in size ranges from 36 in. d i m to 4 ore concentrates, as well as for many of the nonmetallic minerals.
or 5 in. diam. The 4- and 5-in. s i m when used in high-impact mills To obtain a reasonably low pebble wnsumption per ton of material
such as semiautogenous mills may require a special tempering treat- ground, the mills should be designed to operate at relatively low
ment to reduce their hardness and impmve their toughness. The cast s p e d , with low-lift liners, and preferably with a high pulp density.
steel balls are regularly produced in the 141- to 4-in. size range, High-alumina ceramic balls have steadily increased in popularity
which somewhat restricts their field of application. since their introduction. They are normally suited for use under the
The type I highchromium martensitic iron balls are about as same conditions as flint pebbles. However the ceramic balls have
resistant to breakage, though not quite as resistant to spalling, as the advantage of higher spccific gravity and less susceptibility to spall-
the type 3 martensitic steel balls. They have outstanding abrasion ing and breakage. While the cost per pound of ceramic balls is roughly
resistance when grinding minerals softer than about 6.5 Moh's hard- ten times that of flint pebbles, the consumption of ceramic balls per
ness. In grinding cement clinker, these high-chromium iron balls have ton of material ground can, under favorable conditions, be less than
from three to seven (and possibly ten) times the abrasion resistance one-tenth the consumption of flint pebbles.
of martensitic steel balls. Consequently they are now extensively used Sintered tungsten carbide balls, which are made from finely ground
in cement mills, even though their first cost may be two to three tungsten carbide bonded with about 6% cobalt, have outstanding
times higher than the cost of the steel balls. There are also indications abrasion resistance, combined with good resistance to spalling and
from small-scale marked-ball wear tests that these highchromium breakage. Their high specific gravity, which is about twice that of
iron balls can be economically justified in fine-grinding hematite ores steel, indicates that a tungsten carbide ball charge of a given volume
prior to pelletizing.' However in wet-grinding high-silica ores, the should produce about twice the grinding rate of an equal volume of
high-chromium imns are only 25 to 50% more abrasion resistant steel balls. The tungsten carbide balls are normally available only
than the martmsitic steels, which is insufficient to provide a favorable in small sizes, ranging from % to % in. d i m . The very high first
cost-wear ratio. The high-chromium iron balls are available in a W- cost of tungsten carbide balls, which is about 100 times that of steel
to 4-in.diam. size range. on a weight basis, restricts their use to special-purpose applications
Ni-Hard balls (type 2 in Table 3) have been used quite extensively in small grinding mills.
in cement mills and in some oregrinding mills. Their use has been In considering the future for potential materials for grinding me-
largely confined to low-impact mills, since they are more susceptible dia it appears unlikely that there will be much further improvement
to breakage and spalling than the types 1 and 3 wmpositions. Ni- in the steel compositions, since the steel grinding balls and rods are
Hard balls have excellent abrasion resistance in cement-finishingmills, currently approaching the maximum hardness and near-optimum car-
though not as good as the high-chromium iron balls. In recent years, bon content attainable. The type I high-chromium iron composition
the production and use of Ni-Hard balls have declined due to increas- may possibly be further improved in abrasion resistance or alternately,
ing competition from the types I and 3 compositions, which offer its cost relative to steel may be lowered. Further improvement of
the advantages of better toughness and better cost-wear ratios. high-alumina ceramics may be possible. One other approach to the
The type 4 wmpositions, which cover a number of abrasion-resis- production of super-quality balls may involve the use of powder-metal-
tant irons still in the experimental or commercial development stage, lurgy techniques in which hard. fine-grained compounds are bonded
deserve consideration by users of grinding balls. While only a few together by a suitably tough matrix. This is the technique used in
of these are likely to survive the test of time and competition, their making sintered tungsten carbide balls, in which the hard carbide
production and use stimulate the development of more abrasion-resis- is bonded in a cobalt matrix. However the possibility of using other
tant and more economical balls by all of the ball producers. lowerast hard compounds, bonded in a lower-cost matrix, has been
The type 6 wmpositions, representing the unalloyed or low-chro- largely unexplored. It is reasonable to expect, therefore, that further
mium pearlitic white irons, are only suitable for use in low or moder- improvements in the composition and quality of grinding media will
ate-impact mills. They were extensively used until about 1950. Since continue. It will, however, be necgsary to balance such improvements
that time, their rapidly increasing production cost, their relatively against cost of production, since practically all grinding media for
poor abrasion resistance, and the trend toward use of larger higher- high-tonnage operations are selected on the basis of their cost-wear
impact mills have all combined to reduce the popularity of these ratio.
pearlitic white-iron balls.
In general, the cost of producing all types of balls has increased Grlndlng Mlll Llners
rather rapidly since the end of World War 11. However it does not While the principle of grinding ores and minerals with tumbling
cost much more to produce highquality balls, such as the type 3 media was recognized at least as early as the mid-19th century, the
composition, than it does to produce the less-abrasion-resistant types development of mechanically suitable mills to grind in this manner
5 and 6 compositions. Consequently the economic choice of ferrous- did not occur until about 1908. Following their development. these
alloy balls appears to be narrowing down to the type 3 composition mills were rapidly adopted to replace Chilean mills and other methods
for wet grinding of siliceous ores and hard minerals, and the type I of grinding. It quickly became obvious that these mills would have
composition for grinding the softer minerals, such as cement clinker to be lined with a suitable abrasion-resistant material, since the struc-
and possibly iron-ore concentrates. As previously mentioned, however, tural grade of steel or gray iron used in the shell and ends of these
one or more of the experimental compositions in the type 4 classifica- mills tended to wear out quite rapidly.
tion may also prove to be economicalIy competitive. During the period from 1908 to about 1912, when much progress
In selecting grinding media from the nonferrous materials, the was made in the development and design of grinding mills, the choice
principal candidates are flint or other hard, fine grained rock, high- of suitable materials available for mill linings was quite limited. Cast
alumina ceramic balls, and sintered tungsten carbide. Ore lumps, as white iron, Hadfield-type austenitic manganese steel, and unalloyed
used in autogenous and semiautogenous grinding, are also an impor- high carbon steel were the principal ferrous materials suitable for
3G14 GRINDING
these linings. A nonmetallic lining, made from cut flint blocks or a range of compositions, containing 14 to 23% chromium, which
natural stone, which were cemented-in to form a smooth Silex lining, also contains sufficient molybdenum plus optional amounts of nickel
was also used quite extensively in fine-grinding mills, which generally or copper so that heavy section liners of this alloy can be hardened
used flint pebbles for grinding media. There was also considerable by heat treatment to produce a structure of chromium-iron carbides
use of composite linings made from highcarbon steel and concrete, (CrlC3 type) in a hard martensitic matrix. This alloy iron normally
or the use of liner designs which would trap and hold grinding balls has the best abrasion resistance of any of the ferrous materials listed
or pebbles in pockets, to form part of the wearing surface of the in Table 13. It also has surprisingly good resistance to spalling and
liner. Most of these early mills had relatively small diameters, usually breakage in service. There are, however, some high-impact ball mills
in the range of 3 to 7 ft, which did not impme the requirements on and rod mills which will break or spall liner plates of this material,
materials for liners which are demanded by the largediameter ball so its use should be approached with caution where high-impact condi-
and rod mills used in many present-day grinding operations. tions are known to exist. Also, the alloy normally carries a higher
The trend toward large-diameter mills began in about 1926 when first cost than any of the other materials listed in Table 13. Conse-
9-ftdiam mills were installed at Andes Copper Mining Co. Similar quently, it may be economically competitive for use in one operation
mills were installed a few years later at Climax Molybdenum Co., but not in another.
followed by 10-ft-diam. mills at the Morenci plant of Phelps Dodge Item 2, representing several grades of higharbon martensitic
Corp. From about 1950 up to the present time, when many new steels, contains sufficient chromium and molybdenum, plus optionally
mills were installed to grind taconite iron ores, low-grade copper some nickel, so that heavy section liners made from these grades
ores, and also cement plant feed and products, the trend toward large can be heat treated to full hardness by a quench in air or molten
high-horsepower mills continued, so that at present the maximum salt. A grade containing about 1% carbon, 5 to 6% chromium, and
size of rod mills is about 15-ft diam and maximum size of ball mills 1% or more molybdenum, which is similar to the air-hardening type
is presently about 18-ft diam. It is obvious that shutdown time for of A-2 tool steel, has excellent abrasion resistance plus good spalling
relining these large mills should be kept at a minimum, so there resistance, but tends to develop high internal stresses when produced
has been a steadily increasing demand for liner materials with high in heavy or nonuniform cast sections. Consequently, it has been some-
abrasion resistance. what prone to cracking in the foundry or during service. This has
Between about 1930 and the present, there have been literally limited its commercial use, which at present is largely confined to a
hundreds of ferous materials investigated or tested for mill liners. number of cement milling operations in Europe. Another grade in
A few of these could be classed as new alloy types, while many others the item 2 classification contains about 0.7% carbon, 1% manganex,
represented variations in composition, heat treatment, hardness, and 1.5% chromium, and 0.5% molybdenum. It is hardened by a quench
structure of the respective alloy types. In addition, two nonmetallic in molten salt, then tempered to a hardness of about 500 to 550
materials, specifically special rubber formulations and high-alumina BHN to suit operating conditions in the mill. It is less prone to cracking
brick, have been developed for use in mill linings. than the 5Cr-IMo grade, but is also somewhat less abrasion resistant.
The steels and irons presently used in liners for rod, ball, and It has been used quite succgsfully in a number of high- and medium-
pebble mills may be classified into ten general types. The composition impact ball mills in the United States.
and hardness ranges of each type are listed in Table 13. These ten Item 3 represents the high-chromium white iron which was origi-
compositions represent the principal types of ferrous materials which nally developed for heat and corrosion-resistant castings and was
have survived the test of time and commercial use. They are listed then found to have high abrasion resistance, especially when used
in their approximate order of decreasing abrasion resistance, though in slurry pumps. It has found limited use in grinding mill liners in
it should be recognized that minor reversals in this order may occur, the United States, and has now been almost entirely displaced by
depending on character of the mill feed and on operating conditions. the Cr-Mo irons in item I. However, there is still some use of the
The martensitic Cr-Mo white iron, item I in Table 13, is basically item 3 irons for cement mill liners in Europe.
The good abrasion resistance of rubber depends on its ability to occurs by a micro-spalling mechanism. This wear can be materially
yield elastically without being cut or scratched when an abrasive reduced in wet grinding by maintaining a high-viscosity, high-density
panicle impinges upon or is pressed into its surface. Consequently, pulp in the mill and by operating the mill at relatively low speeds
rubber liners tend to provide their best service when the abrasive and with peripheral speeds under 270 fpm. If favorable conditions
panicles are small, as in fine grinding and when the panicles are can be maintained in the mill, a ceramic lining may last 10 times
relatively free from hard sharp corners or edges. Extensive research longer than metal or rubber linings and 20 to 30 times longer than
and field testing have been conducted by the producers to expand natural stone. On the other hand, if these favorable conditions are
the range of operating conditions in which rubber liners will perform not or cannot be maintained, wear will occur by a micro-spalling
successfully. As a result, there has been a steady growth in the use mechanism and the life of the ceramic lining may be substantially
of rubber mill linings, staning in 1961. less than that of a metal or rubber lining in the same mill.
Ceramic linings for fine-grinding mills reprmnt an evolutionary Cast of Liner Materials. Cost of liners is a very important consid-
development from the natural stone linings used in many of the fine- eration, especially in the milling of low-grade ores or relatively low-
grinding mills, particularly pebble mills, during the early days of priced materials such as cement. Total cost of a lining should include
grinding mill development. The natural stone, or Silex linings, were the cost of the liner material, the cost of installation, and at most
usually made from cut flint blocks cemented in place. The development operations, the cost of shutdown time for relining, which in turn is
of ceramic brick to replace natural stone introduced several distinct measured in terms of lost production. The cost of lost production
advantages. The presently available high-alumina ceramic shapes have varies considerably, depending on conditions at each milling opera-
much better abrasion resistance, better resistance to spalling, better tion, but may, under some circumstances, total more than 50% of
strength and toughness, better resistance to heat and chemical attack, the installed cost of the liner. The installed cost of the liner is usually,
and better adaptability to economic production and installation under however, the major factor involved in its selection. The installed cost
presentday conditions. In some cases, these ceramic linings can eco- of liners has increased rapidly over the past 30 years. In terns of
nomically replace metal and rubber linings. Where contamination relative costs, there is a trend toward narrowing the differences be-
of the ground product by wear of metal or rubber linings must be tween the various candidate materials.
avoided, the use of a ceramic lining will normally solve the problem. Table 14 lists typical relative costs of the various candidate or
The production of white cements and titanium paint pigments consti- readily available materials for linings in the western United States.
tute two of the largest markets for ceramic mill liners and also for For the preparation of Table 14, the delivered prices of materials
ceramic grinding media. for shell liners were obtained from the major suppliers of liners in
The original ceramic linings were made from porcelain, which this area. Their prices were averaged, and the approximate cost of
is a fired clay, quartz, and feldspar mixture with a lower alumina bolts for an average lining in mills 7 to 14 ft d i m , plus the estimated
content than the ceramic bricks and shapes now used for mill liners. cost of labor for liner installation, were added to obtain the installed
Ordinary porcelains do not have as good abrasion resistance or spalling cost per pound. To obtain the relative weights, it was assumed that
resistance as the high-alumina ceramics. For commercial use in mill for each mill, the relative volume was the same for each of the ferrous
liners, the ceramic bricks contain about 85% alumina, though for alloy and ceramic linings. A similar assumption was not possible
special applications even higher alumina contents may be used. The for rubber linings due to their differences in design and to the fact
bricks are fired at about 2700°F to provide a high-density, nonporous that the rubber linings actually consisted of rubber plus metal rein-
product. They are normally cemented in place using portland cement, forcement. Consequently, the weight of rubber linings, relative to
acid-resistantcements, epoxy cement, or silicone rubber cements. Sup- the weight of metal linings, was based on data from a number of
plementary fastening with bolts or other types of fasteners may also large mills where rubber linings, as recommended by the manufac-
be used. Special shapes are sometimes used to form lifter bars in turer, had been used to replace metal linings. Where the rubber linings
the mill lining. The high spalling resistance of ceramic bricks makes were approximately equal in volume to the metal linings, their weight,
their use in lifter bars feasible for many fine-grinding mills. including bolts, was between 0.22 and 0.24 of the weight of the metal
Ceramic lining brick has a specific gravity of 3.4, which is about lining.
44% of specific gravity of steel. Hardness of the ceramic is 9 on To obtain the relative operating cost or ewnomy of each lining
the Moh's scale, which indicates it should not be scratched or worn material in Table 14, the relative installed cost should be divided
by the abrasive minerals such as quartz. Wear of the ceramic probably by its hours or days of Life in service. For the ferrous materials.
Tabla 14. Typical Ralativa Costs of Matarials Used for Linings in the Wastarn USA--Dacambar 1973.
Relative Relative
Installation Installed weight of installed
Cost cost per cost wiplete cost of Toughness
ltemt Material per Ih Ib$ per Ib lining lining# ratingv
1 Martensitic Cr-Mo white iron 0.38 0.04 0.42 100 42 6
4 Martensitic Ni-Cr white iron 0.30 0.04 0.34 100 34 7
5 Martensitic Cr-Mo steel 0.36 0.04 0.40 100 40 5
6 Austenitic 6Mn-1Mo steel 0.32 0.04 0.36 100 36 3
7 Pearlitic Cr-Mo steel 0.28 0.04 0.32 100 32 4
8 Austenitic 12Mn steel 0.30 0.04 0.34 100 34 2
10 Pearlitic white iron 0.22 0.04 0.26 100 26 8
R Rubber 1.17 0.10 1.27 29 1
C Ceramic 0.31 0.18 0.55 445 24 9
' Typical costs and weights are based on averaged data from a number of large oremilling operations where comparisons
of the various materials have been or are being made.
t Items 2, 3, and 9 in Table 13 are omitted from Table 14 due to lack of comparative data.
$Installation wst includes cost of bolts plus installation labor, except on rubber, where bolt cost is included in cost per
pound, and on ceramic where cement was used in place of bolts.
(Relative weight of rubber lining includes weight of bolts and metal reinforcement.
5 Ceramic assumed to have same lining volume as the ferrous materials.
$C Relative installed cost of lining does not include cost of lost pmduction during shutdown for relining.
n Toughness rating refers to resistance to spalling or breakage as determined by service experience. Highest order of
toughness = 1 and lowest = 9.
.
TUMBLING MILL CHARACTERISTICS 3617
this service life should be greatest for item 1 and least for item 10.
For the rubber and ceramic liners, the mill operating wnditions have
a strong influence on service life. For fine grinding under favorable
conditions, the rubber and ceramic liners may have substantially
longer life than any of the metal liners, while for coarse grinding of
hard ores or minerals with rods or balls larger than 3 in. diam, the
life of rubber or ceramic linen may be too short to permit their
economic use. Other factors, such as relative resistance to breakage
or spalling, must of course also be considered in the selection of
materials from the list in Table 14. The toughness rating of the materi-
als in Tabk 14 indicates their order of resistance to breakage and
spalling in high-impact service. While this rating is somew!~atarbi-
trary, it does represent a consensus of field experience with the various
materials, when used in liner service.
Development of Liner Designs. In the early days of grinding-
mill development, the liners were designed primarily to permit the
most economical use of the available liner materials. Not much atten-
tion was paid to the effect of liner design on the grinding efficiency 2 I".
or grinding capacity of the mill. However, it soon became apparent
that there was a close relationship between overall mill performance,
2:: in.
31" ,
including grinding media consumption, and such factors as liner thick-
ness; the lifting effect of waves or lifter bars; the spacing, height,
and shape of the waves or lifters; and the wear pattern which developed
on the liners during their period of useful life. It was found that
liner designs which worked well in one mill and one condition of
operation might be quite unsatisfactory in another mill having a differ-
ent size, speed, grinding media, or another wndition of operation.
Consequently, in spite of the many years during which innumerable
liner designs have been investigated and used, there is no one design
which has been found to give optimum performance in all mills.
Liner designs are also influenced by their cost of production. Cer-
tain cast designs are more difficult to cast or heat treat than others
or may require more man-hours per unit weight of metal cast. If f.
rolled steel liners are used, the user is restricted to certain standardized FIG.I. Rod-mill shell line% typical c r m uctions.
cross sections, since it is impractical for a steel plant to produce
special shapes or special rolls to suit the individual preferences of which include mill speed and diameter, size of rods used, abrasiveness
each milling operation. and size of the mill feed, discharge level of the pulp, and toughness
For rod, ball, and pebble mills the design, or more specifically of the material in the linen. There is no one liner design which is
the cross-sectional contour. is of most concern in the shell liners, test for all mills. However, the single-wave design shown in Fig. la
since their thickness, lifter or wave height and spacing, and wave is currently the most popular for heavy-duty service in the modem
contours have a definite influence on liner life, efficiency of utilization large-diameter mills. When installed, it is typically about 3 in. thick
of the liner material, mill grinding rate, and degree of impact experi- in the valleys and 6 to 7 in. thick at the top of the wave.
enced by the grinding media. End-liner designs, on the other hand, Designs a through e may be produced as castings, which may
are selected mainly on the basis of liner life and metal utilization. be made from items I through 8 of the materials listed in Table I.
They may be varied in thickness along their radial dimension in order Designs la and le are also produced as rolled sections made from
to obtain a low scrap loss on the liner when it is worn out. Rod- items 5 and 9 in Table 13. Where two-piece designs such as Id and
mill end liners usually have a smooth inner surface so that end lift le are used, a relatively tough material such as cast or rolled alloy
of the rods is avoided. Ball and pebble mill end liners preferably steel may be used for the lifter or clamp bars, while one of the less-
have lifter bars on their original wearing surface to reduce wear and tough but more abrasion-resistant alloy white irons (such as items
extend the liner life. Discharge grata, which also act as end liners. I and 4) may be used for the plate sections.
preferably have lifters or lifter-type clamp bars on the wearing face The abrasive conditions which exist in most rod mills do not
of the grate assembly to reduce wear and extend the life of the grates. favor the use of rubber linings, while the impact conditions practically
Shell-Llner Design for Rod Mills. Rod mills provide an interme- rule out the use of ceramic linings. However, for certain rod mills
diate step in ore sizereduction between the crushing plant and the grinding the softer ores or minerals, some degree of success has been
ball-milling operation. W. Jurden16 has outlined the history of their achieved with the use of rubber linings. Most of these mills are of
development. With the growth in use and size of rod mills after World relatively small diameter, which may be a factor in the successful
War 11, it became apparent that the design of the shell liners had a use of rubber linings in rod mills. The design of the lining is also
marked influence on both liner life and proper functioning of the an important factor. The height, spacing, and design of the lifters
rod charge. A symposium on rod-mill liner design." which was held is particularly important. A typical design for rubber rod mill liners
at the American Mining Congress in 1950, discussed a number of is illustrated in Fig. If.
these design problems. A modern large-diameter rod mill may grind ore at a rate of
For the present-day commercial-size rod mills, it has been found 100 to 200 tph. Obviously shutdown time for relining is an important
that the shell liners should have waves or lifter b a n to produce a consideration. One method of reducing this time is to increase the
definite lifting effect on the rod charge. When this lifting effect is length of individual liner plates, so that fewer plates are required
lost as a result of liner wear, the lining tends to wear out quite rapidly for a complete lining. For example, at the Bonneville plant of Kenne-
and the grinding rate of the rod charge also drops. Some of the cott's Utah Copper Div., shutdown time for liner replacement was
more successful sections which have been used for rod-mill liners substantially reduced by increasing the length of each liner plate so
are illustrated in Fig. I. The choice of a particular design from the that only three liner plates were required per row, in an 1841-ft-
six basic types shown in Fig. I is influenced by numerous factors long mill, in place of seven plates in the original design.16
Shell-Liner Design for Ball and Pebble Mills. For purposes of
identification and discussion in this section, ball mills include all mills
which use a tumbling charge of metal balls, cylinders, or conical
shapes to grind ores and industrial minerals. This does not include
autogenous mills where large p i e m of ore or rock are used as grinding
media, or semiautogenous mills, where a mixture of large pieces of
ore or rock and relatively largediameter balls is used for grinding.
Pebble mills, as discussed in this section, include those mills which
use natural stone or flint pebbles or ceramic balls for fine grinding
of ores, concentrates, or other materials such as cement clinker. Where
the pebbles are obtained by screening them out of the ore being pro-
cessed, as in fine grinding of gold ores in South Africa, there is no
clear distinction between pebble-mill and autogenous grinding. How-
ever, insofar as mill liners are concerned, no particular distinction
needs to be made between liners for nonautogenous pebble mills and
those for autogenous pebble mills.
In ball mills for primary grinding of crushing-plant or rod-mill
products, the balls used for makeup normally range from a maximum
of 5 in. down to about 141 in. diam. Currently the 5-in. size is seldom
used. By far the greatest tonnage of balls for primary grinding is
now in the 3- to 2-in. size range. For fine grinding, which includes
secondary grinding of ores, concentrates, and other materials such
as cement, the ball sizes range from about 2 in. down to the minimum
commercially available size, which until recently was about W-in. e. I
diam. During the past few years, there has been some production i.
of cylindrical-type grinding media, with diameters and lengths down FIG.2. Ball and pebble mill shell linen, typical cross sstions.
to about %-in. size. These Minipebs are used for very fine grinding
of rapid-hardening portland cements. When pebbles are used in pebble of new liners is 6 in., then its relative grinding rate is 223. If the
milling, they usually have nominal diameters ranging from about 4 average thickngs of new liners in this mill is 3 in., then its relative
in. down to about 2 in. Ceramic balls range from 2-in down to %- grinding rate is 261, which represents a 17% increase. This improve-
in diam. ment will be somewhat offset by the fact that there will be more
There is a fairly definite correlation between the size of balls in shutdown time for liner replacements when using the thinner linings.
a mill and the design used for linen. The larger sizes of balls are Also, on the basis of experience at a number of milling installations,
used for coarse mill feeds, which tend to produce high wear rates the 6-in.-thick liner would utilize about 75% of its original weight
on the liners. Consequently, the liners are usually made thicker and by the time it was worn out, while the 3-in.-thick liner would utilize
with higher or more waves or ribs to extend their life. In fine or only about 60% of its original weight. Obviously a compromise be-
secondary-mill grinding with small balls, the wear rates on the liners tween the thickness of new liners and the other factors governing
are slower, which in turn permits the use of thinner liners with lower operation of a particular mill is necessary. At present the average
waves or fewer liften. The smaller balls also produce less impact, thickness of new liners in ore-grinding mills is about 4 % in., though
so that more brittle materials may be used in the liners. where maximum throughput is a prime consideration, the average
When liner plates are made from castings, the possible variations thickness of new liners may be held down to about 3 in. In cement
in their design are infinite. Consequently, there have been literally mills, where wear rates are low, the new-liner thickness may be held
hundreds of different designs tested over the past 65 years. Taggart" down to about 2 to 3 in.
provides illustrations and discusses some features of the more success- Where a plate and lifter-bar type construction is used in ball mills,
ful designs which were developed up to about 1943. Some of the the height of the lifter bar is an important consideration. A combina-
designs which he shows are no longer in use, for various reasons, tion of high lifters and small balls will produce a dead space filled
while several new designs and new materials, which have been devel- with inactive balls in front of each lifter bar, thereby reducing the
oped during the past 30 years, are now used extensively for ball- effective mill diameter. Experience has indicated that for secondary
and pebble-mill liners. Fig. 2 iIlustrates the more s u ~ s f u cross-l or fine grinding with small balls, the lifter heights should be lower
sectional designs of shell liners made from cast ferrous materials. than for primary or coarse grinding with large balls.
Cast linen, while still very popular, are now faced with increasing The maximum length of liner plates is limited to the length which
competition from rolled steel, rubber and ceramic liners. On the other can be conveniently cast in the supplier's foundry. and by the length
hand, the cemented-in railroad rail linings, as described by Taggart," of the mill. Normally for convenient installation a minimum of two
which were used in a number of large oreancentrating operations liner plates per row is necessary. The use of two liner plates per
during the 1930s and early 194% have now bem replaced, with one row will require less installation time than three or four liner plates
or two exceptions, by cast ferrous-alloy or rubber liners. per row. For example, in the 13 x 12-ft ball mills at Climax Molytde-
The maximum width of the Liner plates in Fig. 2 is limited to num Co.,a change from four 3-ft liners per row to two 6-A liners
the size which will go through the manhole opening or discharge per row reduced installation time for a full set from about I2 hr
trunnion of the mill. Actual widths are usually in a range of 15 to down to a range of 6 to 7 hr.' This experience is in line with Kenne-
24 in., which covers a I2 to 20° segment of the shell's circumference. cott's e~perience'~ on rod-mill shell liners. A 6- to 7-ft liner length
The average and maximum thickness of the liners is limited by several is approaching the practical maximum for production at most foun-
factors. The producing foundries object to liners which are more dries. The circumferential joints in each row of shell liners should
than 7 to 8 in. thick at the top of the waves, since with increasing be staggered to prevent excessive grooving by wear. This can be done
thickness it bsomes more difficult to heat treat them to the desired by using two or more dikrent lengths of liner plates in the assembly,
structure or hardness, and it is also more difficult to avoid a shrinkage or by designing the liners so that the joints (except at each end of
cavity in the center of the casting. Thick liners also reduce the grinding the mill) form an angle with a circumferential line around the mill.
capacity of the mill, especially on those of smaller diameter. The In Fig. 2, saction a illustrates a block or flat-rib type of liner
grinding capacity of a mill increarcs as the 2.6 power of the mill used in some primary ball mill and rod mill liners. Height of the
diameter. If, for example, on a 9-ftdiam mill, the average thickness rib is adjusted to give the desired lifting effect and may vary from
TUMBLING MILL CHARACTERISTICS 3619
a minimum of M in. in ball mills to as much as 3% in. in some tests in 13 x 12-ft mills at Climax, operating at about 67% critical
rod mills. As this design wears in a ball mill, there is a tendency speed, indicate that a high-wave, double-wave design in these mills
for a gouge or depression to form at the leading edge of the rib. will produce a slightly better grinding rate and a substantially longer
thus increasing its lifting effect and at times producing an undesirably liner life than the formerly used shiplap linen. It is indicated therefore
high lifting d e c t on the ball charge. Consequently, this rib-type liner that the percent critical speed of a mill is an important consideration
is not as popular as the single-wave design, Fig. 2b, which normally when choosing a liner design. The results from a number of other
wean down more uniformly. However, if the wave is too high on tests at Climax, which compared the performance of various designs
this single-wave liner, a gouge or depression may also form on its of shell liners in the primary mills, have been summarized by
leading slope which may produce an undesirable ball action and re- Wind~lph.~'
duced grinding rate during a period about midway between the begin- The triple-wave liner, Fig. 2i, was developed at Climax to replace
ning and end of its life. Height of the wave in this single-wave liner thin shiplap liners in the 9 x 9-ft Marcy mills. It is 3% in. thick at
is normally between 3 and 4 in. the top of the waves and 2% in. thick in the valleys. It retains most
The shiplap liners, Figs. 2c and d, are designed to produce a of its original wave contour as it wears away, so there is little change
more gentle lifting effect and more spinning or rolling dfect on the in ball action or lifting effect during the life of the liner. This triple-
balls. They normally maintain a wear pattern which produces the wave liner design has worn at an average rate of about 5.2 Ib per
desired ball action throughout their service life. The Fig. 2c design hr compared to about 7.2 Ib per hr for the thin shiplap liners made
is the preferred choice when the liner is made from an alloyed white from the same alloy steel and used in the same mills. At the same
iron, while the Fig. 2d design is normally a better choice for steel time, a small improvement in grinding rate was indicated by the
liners. The Fig. 2d design will normally wear to a lower scrap weight change to the triple-wave design. The design appears to be well suited
than the 2c design. These shiplap designs were extensively used in to mills operating at high critical speeds, especially when maximum
ball mills for many years. However, during the past 20 yean, many grinding capacity is required from the mill.
mills have changed from this design to the double- and triple-wave Liners with the ribs or waves cast to form a spiral pattern on
designs, as illustrated in Figs. 2h and i. the inner face of the lining have been used at a number of milling
The plate- and clampbar or wedge-bar design, Fig. 2e, can be operations. They can be designed to convey the ball charge towards
made with a tough steel used for the clamp b a n and a harder more the feed end of the mill where the larger-size balls tend to remain
abrasion-resistant material used for the plates. Also it o C r s an advan- and thus produce more rapid or efficient grinding of the coarse parti-
tage, at least in theory, in that two or more sets of clamp bars can cles of feed entering the mill. By reversing the direction of the spiral
be worn out for one set of plates. In actual practice, however, it is to produce a conveying action towards the discharge end, a higher
difficult to match the life of one set of plates to two or more sets of circulating load through the mill is possible. Improvements in mill
clamp bars, especially when they are made from austenitic manganese performance are claimed by both methods of using these spiral liners.
steel. This design has lost its popularity for use in ball mills, though The spiral-type lining costs more to produce, and it tends to wear
modifications of the design are now used in some autogenous mills unevenly to produce a high scrap loss. Consequently, it has not been
and some rod mills. accepted for general use in ore-grinding mills. However, in the rela-
The single-wave design, Fig. 2f, with its low wave, is seldom used tively long mills used for grinding cement clinker, where segregation
in primary ball mills due to its tendency to wear rapidly and its of the largerdiameter grinding media towards the feed end is a definite
lack of lifting effect on the ball charge. It is used in secondary or advantage, the use of linings with spiral liften can in some instan-
fine grinding mills where wear rates are low, and is particularly useful replace chamber grates as a method of maintaining segregation of
in pebble mills, where a low lift is desired to avoid breakage or spalling large grinding media towards the feed end and the smaller grinding
of the pebbles or ceramic balls. media towards the discharge end.
The Howes liner design, Fig. 2g, was developed and patented Another liner design, which will produce segregation of ball sizes
by W. Howes9 in the early 1940s. This double-step design was propor- in long mills, is illustrated in Fig. 3. This Slegten lining, as described
tioned to minimize slippage of the balls while in contact with the by Slegten and Mattan.12 is particularly useful in the second compart-
liner. At the same time. the valleys were so designed that the balls ment of cement clinker mills. where it has provided a substantial
did not become wedged into them to form a smooth lining. The increase in grinding rates and reduced kilowatt-hour per ton of cement.
dimensions of the steps and valleys were proportioned to suit the The lining segregates balls by use of the same mechanism which
size of the replacement balls charged into the mill. The valleys were produced segregation of balls in the original Hardinge conical mills.
about one balldimmeter deep and about wide enough to seat one- An advantage of the Slegten lining is that it is readily adaptable to
half of the ball volume therein. Plateaus on the steps were about the present design and bolt-hole spacing of cylindrical mills.
one balldiameter wide. This design produced an improvement in Various designs of mill shell liners have used rolled high-carbon
grinding rate and liner life when compared to the block or shiplap steel sections. Taggartl" shows a number of the early designs, including
l i n m illustrated in Figs. 2a and d. the Osborn design, which was quite popular in South Africa. In the
The good results obtained with the Howes design served to stimu- United States one of the successful and economical liners was the
late interest in new designs for shell liners. In the late 1940s. Broughm rail lining, in which used railroad rails were cut to suitable lengths
at the Bayard mill of US Smelting, developed the double-wave design, and cemented in place to form a relatively smooth lining when new.
Fig. 2h, which gave sufficientlygood performance in hi 7 x 6-ft Marcy As the cement wore away, the grinding media would wedge in the
grate mill to interest other larger operations in this-design. Morenci pockets between the rails to form a part of the lining. Linings of
was probably the first to adopt it for large-scale use and this was this type were relatively low in cost and had surprisingly good life.
followed by many of the other copper concentrators in the southwest- They would probably be unsuitable for largediameter mills but they
ern United States. At present, this double-wave liner is probably the have been quite satisfactory for mills up to 7 A diam. They are best
most popular cast design for primary ball mills. It should be noted, adapted to use where a large number of mills are in operation, so
however, that when the double-wave design, which produces a rela-
tively high lifting effect on the balls, was used to replace the shiplap
design in 9 x 9-ft primary ball mills operating at 78% of critical speed
at Climax, a loss of about 7 to 8% in grinding rate war experienced
on new liners.' Average loss during the entire life of the liners was
about 4%. Liner life was increased almost 60% by the change to
the double-wave type, but the loss in grinding rate in these mills.
which were being pushed for full capacity, ruled out the further use
of the double-wave design in these higher-speed mills. More recent FIG. 3. lntmul vim of cornparlrnenl mill with Slegtm lininq.
3620 GRINDING
The S-100 and S-200 linings shown in Fig. 4 are of the relatively
new one-pi= type. They are rolled to the cross-sectional contour
shown, then cut to the desired lengths; the edges are machined to
provide the desired chord width and each plate is heat treated to
hardness above 500 BHN. A special high-hardenability alloy steel
(item 5 in Table 13) is used for these S-100 and S-200 sections. The
Type "A" linings T S-100 design is 3 in. thick in the valleys and 641 in. thick at the
Type "D" linings top of the wave, which makes it well suited for rod mill liners.
I The S-200 double-wave designs are either 2 in. thick in the valleys
and 4 in. thick at the top of the waves or are 241 in. thick in the
valleys and 5 in. thick at the top of the waves. They are specifically
designed for use in ball mills. The various modifications of the US
Steel rolled-section designs make them adaptable for use in mills of
both small and large diameters. They are used under a wide variety
, of conditions, ranging from highly abrasive high-impact rod mill and
Type "B" linings primary ball mill operations to mildly abrasive cement grinding with
"5-100" linings
small balls in ball mills.
Since about 1961 there has been a marked growth in the use of
rubber linings throughout the world. Much of this growth is due to
the development of successful designs by the rubber manufacturers.
The principal use of rubber linings has been in fine-grinding opera-
tions, where they are capable of providing excellent wear life. The
"s-200" linings liners generally involve the use of molded plates wedged between
Type "c" linings molded lifter-bars, as shown in Fig. 5. Channels or slots arc molded
FIG.4. Rolled-steel shell linen, typical crm scctions (US Steel Corp.) into the lifter-bars to provide suitable fastening systems. which vary
somewhat in design, according to the preferences of each manufac-
that a trained full-time lining crew can be justified. Due to the rela- turer. The manufacturers of rubber liners are prepared to make de-
tively long time required for each relining job, it is necessary to have tailed recommendations on the design of liners to suit each particular
one or more spare mills, so that a mill with a worn-out lining can milling operation. This includes such variables as plate thickness and
be removed from its operating stand and taken to a repair bay for width, the dimensions and number of lifters, and wave contours. To
relining. In recent years, the trend toward large-diameter mills, the some extent, the choice of designs for a specific lining is limited to
shortage of scrap rails, the desire to use thinner wave-type liners the mold designs which the manufacturer has on hand. However,
for increased grinding r a t s , and other factors have made the use of where it appears necessary or advisable, the manufacturer may agree
the rail-cement lining unattractive for most operations. It is, however, to make up new molds to suit the ngds of a particular installation.
still used in the 7 x 10-ft ball mills at the Magna and Arthur plants Ceramic linings, as illustrated in Fig. 6, are made from high-alu-
of Kennecott Copper, where a total of about 120 mills are in service. mina ceramic bricks 4 x 6 in. on the wearing face and are usually
In about 1939, the US Steel Corp. introduced their LONUrolled- I !4 or 2 in. thick. Brick 4 in. thick can also be supplied. They may
plate lining. Fig. 4 illustrates the evolutionary and most recent designs be installed without lifters to produce a smooth lining, or courses
for this lining. The original design, type A (Fig. 4), consisted of rolled of lifter-bar brick may be installed at any multiple of the 4-in. brick
high-carbon steel plates, 1 to 2 in. thick, curved to fit the mill shell spacing, to obtain a 141- or 2-in. lifting effect on the balls or pebbles
and held in place by wedge-lifter bars. The type B, C, and D linings used in the mill. The bricks are cemented in place with portland
(Fig. 4) represent more recent modifications of the two-piece design cement, epoxy resin, or silicone rubber. Specially formulated acid-
which are adaptable to plates 2 to 3 in. thick. The lifter bar projects resistant cements can also be used where n-ry. Cost of labor
1% to 2!4 in. above the plates. The edges of each plate are machined plus cement for installation of a 2-in. thick lining is normally estimated
to provide a close-tolerance fit with each lifter bar. to be 50% of the cost of the brick.
Type of lining:
el- Standard
I ; ! : double
: wave
Single wave
shiplap
pebble
Table 15. Typical Wear Rates of Mill Linings in Primary Grinding of Various Ores
Liner wear, x lb per ton Liner wear, Ib per kw-hr
Plant Unit size,
No. Mill feed. in. ft Rod mills Ball mills Total Rod mills Ball mills Combined
Single Stage Ball Milk
1 Copper ore, -%A 10 x 10 grated
2 Copper ore, -% 12 x 12 grated
3 Copper ore, -% 10% x 11 o'flow
4 Copper ore, -% 16% x 19 o'flow
5 Molybdenum ore, -% 9 x 9 grated
6 Molybdenum ore, 13 x 12 o'flow
7 C~pperore, -% 101/2 x 12 grated
8 Molybdenum ore, 13 x 12 o'flow
' Per ton of feed to rod mills. Part of silica removed between rod mills and ball mills
in plant 5. However the 13-f?mills had a slightly higher liner consump- listed in Table 15, the service life of the liners in the single-stage
tion. This was due to higher consumption of end linen. In the 13- ball mills ranged from a minimum of about 200 operating days to
ft mills the end linen accounted for 44% of the total wear, while a maximum of 500 days. For the double-stage grinding operations,
in the 9-ft mills the feed end liner plus discharge grate assembly the rod mill liner life ranged from 300 to 500 days while the ball
accounted for only 32% of the total wear. The comparison points mill liner life ranged from 900 to about 2,000 days. These Lives were
to the fact that as mill diameter increass, end-liner wear becomes influenced mainly by the abrasive characteristics of the ore, the abra-
an increasingly important consideration. This is probably due to the sion resistance of the liner material, and to some extent by the design
higher impingement velocity of balls against the end liners in the and thickness of the new liners.
larger-diameter mills. The grinding mills used at cement plants probably constitute the
A comparison between liner wear in single-and double-stage grind- second most important application for mill liners. Liner wear per
ing in a rod mill plus ball mill combination would produce higher ton of cement produced tends to be substantially less than in ore
total wear rates. However, the data in Table I5 do not support this grinding, since the cement clinker and the raw materials ground in
conclusion. If anything, the total wear rates on liners are somewhat cement plants are usually much less abrasive than the low-grade high-
lower in the double-stage grinding operation. One principal reason silica ores. The dry-grinding practices in cement plants also favor
for this is that for second-stage grinding in the ball mills which follow low wear rates on the linen. However, these factors tending to produce
rod mills, relatively small balls are used. which in turn produce rela- low wear rates on linen are partially offset by the need for fine grinding
tively mild impact conditions on the liners. Consequently, for the and consequent high energy consumption of both the raw cement
double-stage grinding plants listed in Table 15, the highly abrasion rock and the clinker. Raw milling in cement plants consumes about
resistant martensitic white irons or steels (items 1, 4, and 5 in Table 30 kw-hr per net ton, clinker mills consume about 40 kw-hr per
13) were used in the ball mills. Also, in the case of plants 9, 10. net ton. while the ore-grinding operations listed in Table 15 consumed
and 16. which were grinding the very abrasive taconite ores. the wear from 5 to I2 kw-hr per ton in the one or two stages of primary
r a t s on the rod mill linen were kept low by use of these martensitic grinding.
iron or steel alloys. Recently, it has been found that in most of these Table 16 summarizes the liner wear rates reported by 35 cement
double-stage grinding plants, rubber liners will give good senice in plants, located principally in Europeand in North and South America.
the ball mills, which should bring about a further reduction in total From Table 16 it can be seen that in the grinding of raw cement
liner wear and liner costs. rock, the liner wear rates in wet-grinding mills are from five to seven
In addition to liner consumption, the service life of liners is an times greater than the wear rates in dry grinding. Also there is a
important factor influencing their overall economy. At the plants wide range in the reported wear rates from individual plants. This
-::
some of the raw mills in cement plants have used rubber liners quite
successfully, though no data on specific wear rates of these rubber -
liners are yet available. Rubber liners are not suitable for clinker 8 70
.- .. '
mills where the high operating temperatures would rapidly destroy
.-
* \\
65-
the wearing qualities of rubber. However, ceramic liners have been
used successfully in some clinker mills and are particularly useful,
in conjunction with ceramic grinding media, for production of white
cement. 0 5 10 15 20 25
As is indicated from the wear rates of liners in Table 16, the Service life. l o 3 hours
service life of grinding mill liners in cement plants tends to extend FIG. 10. Relation of ball s i and
~ critical spcad to renice life of rubber mill
over a wide range. Minimum service life is about one year for wet liners (after Trellebog).
grinding in raw mills, while maximum service life may be ten years
or more for dry grinding in raw mills and finish grinding in clinker Fig. 10, if a life of 5,000 hr is arbitrarily chosen as the minimum
mills. for satisfactory service, then for a mill using 3-in. balls, rubber liners
Selection of Liner Materials. In the selection of a liner material would only be suitable at mill speeds below 75% of critical. Since
for a particular grinding mill, there are so many local factors involved many primary grinding mills operate at speeds higher than 75% of
that only a few general rules or considerations can be stated here. critical, it appears that further improvements in rubber compositions
There are currently three principal divisions in candidate materials or rubber-liner design will be neceosary before rubber liners can be
for liners. These are: (1) ferrous-alloy cast or wrought materials, (2) universally accepted for all primary ball-mill applications. An alterna-
rubber or other elastomers, and (3) ceramic or natural stone materials. tive for new grinding plants might be to design the mills so they
A fourth material, specifically sintered tungsten carbide, has also been operate at speeds below 75% of critical. Also for dry-grinding opera-
used for small special-duty mills, but nonnally its first cost makes tions, it is recommended that the working temperature at the wearing
it economically impractical for most commercial-size grinding mills. surface of the liners should not excad 190°F, though certain special
Future linings may involve the use of bonded composites or laminates grades of rubber may be used up to about 22S°F. As is indicated
made from two or more of the afore-mentioned materials. Ferrous in Table 14, the installed cost of a rubber lining (1974) is somewhat
alloys have been the principal lining material in ball and rod mills. less than that of most of the ferrous-alloy linings of equal volume.
The performance characteristics of these f e m u s materials are well To determine the relative economy of rubber vs. metal or ceramic
known. A suitable composition or type can be selected to meet the linings, the relative cost factor in Table 14 should be divided by
requirements of almost any milling condition, with the possible excep- the expected or obtained service life of the respective materials.
tion of certain mills where contamination of the ground product by Ceramic, porcelain, and natural stone liners may be used in fine-
iron or alloying elements in the liners must be avoided. grinding operations where impact conditions are low so that breakage
The use of rubber or other elastomers for liners provides several and microspalling are avoided. Low critical speeds and high pulp
advantages over ferrous alloy liners in those applications where rubber densities in the mill will substantially reduce the wear rates of ceramic
is technically and economically usable. Some advantages of rubber liners. In most cases, ceramic liners are used in conjunction with
over metal are its lighter weight, better noisedampening properties, ceramic balls or natural stone pebbles. Provided these favorable condi-
reduced installation time, and longer life when used in suitable grind- tions are maintained in the mill, ceramic liners are capable of providing
ing environments. The economical use of rubber liners is largely re- very long service life. They can also be used for dry grinding at
stricted to fine-grinding operations, or to mills which do not have elevated temperatures, where rubber liners would fail quite rapidly.
large sharpcornered particles in the feed, which are capable of cutting High-alumina ceramic brick linings have much better abrasion resis-
into the rubber surface. For example, at the grinding plants listed tance than porcelain or natural stone linings. Consequently, relatively
in Table 16, rubber liners are now giving good service in some of thin high-alumina linings have largely displaced the thicker porcelain
the ball mills at plants 9 through 15, which grind a relatively fine or natural stone (Silex) linings in those mills where ceramic linings
product from the rod mills. Rubber liners also give excellent service are feasible. Principal uses for ceramic linings have been in mills
life in concentrate regrind mills. On the other hand, rubber liners grinding white cement, titanium pigments, and porcelain enamel,
had short service lives and were not economically justifiable in rod- where iron or chromium contamination must be avoided. Ceramic
mill service similar to that at the plants listed in Table IS. To date linings also have excellent resistance to most types of corrosion. In
rubber end liners are not used in rod mills. For single-stage primary some ore-milling plants, ceramic linings may be considered for fine
grinding in ball mills similar to those at plants 1 through 8, where grinding of ores and concentrates, with ceramic balls or pebbles as
balls 3 in. or larger in diameter were present in the grinding media, grinding media. As indicated in Table 14. the relative installed cost
the several tests to date have indicated that rubber liners will not of a ceramic lining (1974) is less than that of metal or rubber linings
provide as good life as femus-alloy liners in this type of service. of equal volume.
However, the economic possibilities for rubber liners in primary grind- In the selection of ferrous alloy linings (including grates) for ball
ing are not yet clearly defined. If the mill f e d is relatively free from and rod mills, there are ten compositions listed in Table 13. However,
hard sharpcornered piecs of rock and the critical speed of the mill there is currently very little production of items 2. 3 and 9 for the
is below about 75% then rubber liners may provide good life and linen used in ore grinding mills in the United States, so these have
be economically justified in primary grinding service. been omitted from Table 14. The remaining seven ferrous alloys in
Fig. 10, from a manufacturer of rubber liners.2'indicates the ser- Table 14 are listed in their approximate order of decreasing abrasion
vice-life expectancy of properly designed rubber mill linings as critical resistance. In selecting a ferrous alloy from this series for use in a
speed of the mill and ball size are varied. In using the curves in particular mill lining, there are three main considerations, namely:
TUMBLING MILL CHARACTERISTICS 3G25
installed cost, relative abrasion resistance or liner life, and relative to a martensitic structure when heat treated in heavy-section liners.
toughness. The relative abrasion resistance of the items varies consid- They are produced in both cast and wrought form. The item 5 steel.
erably depending on service conditions. For example, item I may in addition to providing relatively good abrasion resistance, can fre-
provide 50% longer life than item 7 in primary ball mill liner grinding quently be used in high-impact rod and ball mills where spalling or
a high silica ore, while in rod-mill linen grinding the same o r similar breakage would occur on the item 4 irons and even in some caws
ore, the item I material may have two to three times the life of on the item I irons. Consequently, the item 5 steel is becoming increas-
item 7. Also in ball mill grinding of minerals with a Moh's hardness ingly popular, particularly for shell liners in high-impact rod mills
less than about 6.5, the martensitic irons and steels (items 1. 2, and and in primary ball mills using balls 3 in. or larger in diameter.
5) tend to have from three to eight times better abrasion resistance This steel may eventually displace the less-abrasion-resistanctsteels,
than the austenitic or pearlitic steels (items 6, 7, and 8). while in items 6. 7, and 8. in many applications.
grinding silica or other minerals with a Moh's hardness greater than Item 6 is a modified austenitic manganese steel developed about
6.5, the martensitic irons and steels will normally provide only 30 1959 for use in ball mill liners and discharge grates at the Climax
to 60% longer life than the austenitic and pearlitic steels. This observa- Molybdenum mine in Colorado.s This steel has better abrasion resis-
tion is particularly significant in the grinding of cement clinker, which tance than the Hadfield-type mmganese steel (item 8) along with
may be classed as a soft mineral, where the martensitic irons and adequate toughness for use in high-impact mill-liner and grate service.
steels normally have from three to eight times better abrasion resis- It was also more abrasion resistant than the pearlitic chrome-moly
tance than the austenitic and pearlitic steels. It should be recognized, steels, which it displaced in the Climax mills. It also performed better
however, that martensitic iron and steel linen are more susceptible than the pearlitic chrome-moly steels in discharge grates where its
to breakage or spalling in service than the austenitic or pearlitic steels. combination of gmd abrasion resistance and favorable peening charac-
so this factor must be considered where high-impact conditions exist teristics prevented clogging of the grates by worn balls. This item 6
in a mill. steel has recently been largely displaced by items I and 3 martensitic
The relative wear rates of a wide variety of ferrous liner materials, irons and steels in the 13-ndiam overflow-type primary mills at Cli-
when tested in primary ball mills grinding a high-quartz molybdenum max but it is still the preferred material for the lowdischarge. 9-ft-
ore at Climax, have been determined by a special technique which diam primary mill shell line~sand discharge grates.
used marked 5-in.diam balls having the same composition and struc- The pearlitic high-carbon chrome-moly steels (item 7) combine
ture as the candidate liner materials." These wear rates are listed moderately good abrasion resistance and relatively low cost with ade-
in Table 13. Field experience at Climax and at a number of other quate toughness for use in most high-impact service. They are not
operations indicate that the wear rates in Table 13 correlate quite as abrasion-resistant as the martensitic irons and steels, but are nor-
well with service results on liners in primary ball mills grinding high- mally more abrasion-resistant then the Hadfield-type austenitic man-
silica ores. The differences between the abrasion resistance of the ganese steel (item 8) when used in ball mill liners. Due to their higher
materials, as listed in Table 13, are relatively small. When the same yield strength, chrome-moly steel liners do not flow or warp like
materials are used for rod mill liners, which operate on a coarse the austenitic manganese steel when exposed to the repeated impact
grind, or where minerals saner than quartz constitute the principal from the cascading rods or balls. Consequently, there is less incidence
abrasives, then substantially larger differences in relative wear rates of liner-bolt breakage when chrome-moly steel liners replace austenitic
are usually obtained. manganese steel liners or liner bars. The chrome-moly steels, when
The martensitic white irons, items I and 4 in Tables 13 and 14. made with a carbon content of 0.5 to 0.7% are also well suited for
frequently compete for the same applications and should probably use in ball mill discharge and diaphragm grates. At present these
be considered together. The highchromium-molybdenum white irons pearlitic chrome-moly steels are one of the most popular materials
(item I) were developed during the late 1950s and represent an im- for mill liners, clamp or liner bars, and grates used in moderate to
proved modification of the highchromium white irons (item 3). The high-impact service. However, it should be recognized that this mate-
nickelchromium white irons include Ni-Hard and its modifications. rial may eventually be displaced in many of these applications by
The Cr-Mo irons are normally cast in sand molds then heat treated the more abrasion-resistant alloys, such at items I and 5 in Tabla
to produce a structure of Cr&-type carbides in a martensitic matrix. 13 and 14.
The Ni-Cr irons may be cast in sand molds or, for heavy-section The austenitic 12% manganese steels (item 8) were probably the
liners, their toughness and abrasion resistance can be improved by most popular liner material during the years when ball and rod mills
chill-casting them in permanent molds made from gray cast iron or first came into general use. This grade of steel combines moderately
graphite. These Ni-Cr irons normally receive only a stress-relieving good abrasion resistance with excellent toughness and impact resis-
heat treatment. tance. The wearing surface of the steel tends to work-harden up to
Due to the difference in matrix structure, the item I irons are a maximum of about 500 BHNwhen exposed to the repeated impact
normally tougher than the item 4 irons. The item I irons are also from the grinding media. This work-hardening effwt is accompanied
abrasion-resistant. Consequently, the item I irons can be used success- by some plastic flow and volume expansion of the metal, which may
fully in many mill liners where the item 4 irons would break or produce warpage or high internal stresses in the liner castings. This
spall under the repeated impact of the balls or rods in the mill. How- in turn may cause high stresses and sometimes breakage of liner bolts.
ever, many mills operate under sufficiently mild impact conditions Also, closely fitted liner plates have on a few occasions expanded
to permit the use of the item 4 irons, which have become quite popular sufficiently in service to overstress and crack the shell of the mill.
for use in ore-grinding mills, due to their better cost-wear ratio than The development of better-wearing steels and irons such as items
the pearlitic and austenitic steels. Liners made from items I and 4 I through 7, which are also much more resistant to plastic flow in
irons are best adapted to use in mills where a high pulp level is service, has markedly reduced the use of the austenitic 12% manga-
constantly maintained to provide a cushion between the cascading nese steels for mill liners. At present in the United States, use of
grinding media and the liners. These irons may break or spall when 12% manganese steels is confined largely to certain rod-mill liner
used in shell liners for low-pulplevel mills equipped with discharge applications where the impact conditions are too severe to permit
grates, though they can usually be used successfully in the feed-end the use of the other types of steel and iron. Where the 12% manganese
liners for such mills. The decision on whether to use items I or 4 steel liners are used, it is advisable to specify carbon contents near
irons for a particular mill depends principally on impact conditions the high side of the range in Table 13, since this will improve both
in the mill and the relative cost-wear ratios obtainable from each their yield strength and abrasion resistance, while stiIl retaining ade-
material. Frequently service tests on both items are n-ry before quate toughness for use in mill liner service.
a find decision can be made. The pearlitic high-carbon steels (item 9) have been used principally
Item 5 represents the mediumarbon (0.4 to 0.7%) steels which as rolled bars and shapes. They combine moderate abrasion resistance
contain sufficient alloy content so that they can be fully hardened with adequate toughness and relatively low first cost. In past years,
3626 GRINDING
much of this steel was used in the form of worn railroad rails cut for liners has been toward those which provide best service life, even
to suitable lengths for the mill linings and cemented in place. More though their cost may be a little higher than for an inferior-wearing
recently, at least one major steel producer has made bolted-in liners material. It is probable that this trend will continue. Also it is reasona-
from rolled plates and clamplifter bars in this type of steel. However, ble to expect that through the cooperative efforts of producers and
the present trend is to replace the pearlitic highcarbon steels with users of liners, there will be continued improvements in both the
the more abrasion-resistant martensitic steels and irons, such as items design and quality of materials.
1, 4, and 5 in Table 14. Selection of liner Designs. Many factors involved in the selection
The pearlitic white iron (item 10) was a popular material for of liner designs for a particular mill have been mentioned or discussed
liners in low-impact service during the early days of grinding mill in the previous sections. The fact that some liner designs will produce
development. It was originally sufficiently low in first cost to make better grinding efficiency or grinding rate than others is now well
it economically competitive with the other ferrous alloys in Tables established. For large milling operations, where the grinding mills
13 and 14. in spite of its inferior abrasion and impact resistance. are usually pushed to their full capacity, the use of liner designs
However in recent years, its cost of production has risen more rapidly, which provide best grinding efficiency or grinding rate can provide
on a percentage basis, than the other items in Tables 13 and 14. economic benefits which far overshadow other considerations, such
Consequently, the use of pearlitic white irons for mill liners has practi- as the installed cost of the liners or service Life.
cally disappeared in the United States and also in most other countries. The optimum liner design for a particular milling operation de-
Generally where the white irons can be used without breakage or pends on such factors as mill speed and diameter, character of the
spalling in mill liners, the pearlitic type has been replaced by the mill feed, and on a compromise between the desire for best grinding
martensitic Ni-Cr white iron (item 4) or, in more recent installations, efficiencies vs. best grinding rate. Usually, therefore, it will be neces-
by rubber liners. sary to conduct extensive service tests on the various candidate liner
In considering the overall trends in the selection of liner materials, designs before the optimum design for a particular operation can
the steadily increasing labor rates for liner production and installation be determined. Since it can be expected that innovations in liner
have tended to narrow the difference in cost between the various designs will continue, the service test program is likely to extend
candidate materials. Consequently. the trend in selecting materials through the entire operating life of the milling operation.
2. BALL MILLS
C.A. ROWLAND
Introduction
mineral products used in the dry state. Dry grinding is also used
Ball mills are lined drums, either cylindrical in shape or modified when minerals cannot be dewatered to the required
cylinders that have either one or both ends of the shell, consisting moisture level or when the ground product reacts unfavorably with
of conical sections, that rotate about the horizontal axis. Fig. I I shows liquids. For example, cement clinker must be ground dry.
a cylindrical mill, Fig. I2 a conical ball mill, and Fig. 13 a Tricone Dry grinding requires about 30% more power than wet grinding
ball mill (Hardinge tradename). Steel or iron grinding media, generally for comparable size reducti~n.'~ The total power required in a dry-
in the shape of spheres, are used to grind the ore to the specified grinding ball-mill plant including drying may be double that required
product size. In order to obtain more contact area for grinding and for a wet-grinding plant. Grinding-media and liner consumption in
to simulate the shape of worn balls, balls have been made with two dry grinding reported as pounds of metal consumed per kilowatt-
concave surfaces diametrically opposite each other. Some concentra- hour per ton of orem is 1&20% of that used in wet grinding. The
tors, such as Erie Mining Co., have used slugs cut from worn and Wabush pellet plant, Point Noire. Que.," reported ball consumption
broken rods to supplement the balls in ball mills and save money dropped from 6.3 Ib per ton of ore ground to 2.5 Ib per ton of ore
otherwise lost as rod scrap. Cylindrical and conical s h a p s have been ground when they converted from wet to dry grinding, and a 30%
tried instead of balls, but balls remain as the most common shape increase in power consumption. A number of comparisons made on
grinding media used in ball mills. wet and dry grinding of cement raw materials show metal consumption
Ball mills were a logical development from the earlier pebble mills in dry grinding to be 10% of that in wet grinding.
that used hard natural pebbles such as flint pebbles or sized ore pebbles The capital costs for wet grinding are generally lower than for
(obtained from the ore itself) as grinding media. In the early 1900sZ' dry grinding. When thickening and filtering of the wet-ground product
it was found that when cast iron or cast steel balls were used in are required. dry grinding may have a lower capital cost.
place of flint or ore pebbles, the mills drew more power and gave With opencircuit grinding the ball-mill discharge passes directly
greater production capacity. Advances in technology have resulted to the next processing step without being screened or classified and
in the manufacture of ball mills up to 18 ft diam inside shell, drawing no fraction is returned to the ball mill (Fig. 14). In closedcircuit
up to 8,000 hp. grinding the ground material, undersize, in the ball-mill discharge
Ball mills are employed to grind ores, especially the more abrasive is removed either using a screen or a clssifier with the oversize being
ores, to finer sizes than can be produced economically in other size- returned to the mill for additional size reduction (Fig. 15). The over-
reduction machines such as roll crushers, hammer mills, and impac- size material that is returned to the ball mill is called the circulating
tors. Ores can be ground d r y - d r y grinding--or in a slurry-wet load. Open-circuit ball-mill grinding requires more power than closed-
grinding-using ball mills. Dry grinding nominally refen to less than circuit grinding for products containing similar amounts of top-size
1% moisture by weight. If the moisture content increases by several material. The less the amount of oversize allowed in the product,
percent, dry grinding capacity is significantly reduced as shown in the longer the ore must remain in the ball mill when grinding in
Table 17. The usual range of solids content in wet ball-mill slurries open circuit. This increases the production of extreme fines and thus
is from 65 to 80% by weight. the consumption of more power. The power required for open-circuit
Wet grinding is used to prepare the feed material for unit opera- ball-mill grinding can be estimated using the multipliers listed in
tions such as flotation, magnetic separation, gravity concentration, Table 18 and knowing the power required for closed-circuit grinding
and leaching that require a slurry of liberated valuable mineral and to yield the desired product particle size. For example, assuming
unwanted gangue particles. Dry grindingz7 is employed to produce the desired grind size is 90% passing some specific top size, open-
feed for agglomeration, pelletizing, and pyrometallurgy processes that circuit grinding would require 1.40 times the power to achieve similar
require feed that is dry or nearly so and for finely ground industrial results as closed-circuit grinding.