US2104709A - Closed circuit grinding - Google Patents

Description

Original Filed April 9, 1932 2 Sheets-Sheet 1 NEW MILL FEED FIG. 2.

IIIH] BALL M/LL M/ZL DISCHARGE MI E I SLIME gATER WASHWATER /2 \g I. OVER/ZE j F/N/SHED FINES PLUS FAVOAA LE FRACTION Dl-SCARDED FRACTION OF SL/ME OF SL/ME C/RCULAT/NG LOAD FIG. 1.

V DISCHARGED FIN/SHED PRODUCT Y REC/RCULAI'ED r0 M/LL 1N EN TOR. A R THUR JOHN WE/N/G BY mflclwelw A TTORNEY.

A. J. WEINIG CLOSED CIRCUIT GRINDING Griginal Filed April 9, 193 3 2 Sheets-Sheet 2 FIG. 4. FIG. 5.

INVENTOR. ARTHUR JOHN WEI/W6 (159;; (M; Mud.-

ATTORNEY Patented Jan. 4, 1938 CLOSED CIRCUIT GRINDING Arthur John Weinig, Golden, ol0., assignor to The Mine & Smelter Supply 00., Denver, 0010., The Stearns-Roger Manufacturing 00., Denver, 0010., and The Dorr Company, Inc., New York,

Application April 9, 1932, Serial No. 604,151

Renewed July 18, 1936 12 Claims.

This invention relates to open or closed circuit grinding of materials by a process with the grinding device in circuit with some kind of a particle size classifying or separating means.

The invention has for its principal object to increase the efficiency of the present circuited grinding methods; to lessen the cost of grinding; to increase the capacity of the present types of equipment used in line grinding; to lessen the water used in the circuit; to improve the character of the end product, both of itself and its adaptability forsubsequent treatment; and still other objects are in view as-can be seen from a further reading of this specification.

The nature of this invention resides in results obtained from a searching investigation as to what actually takes place in the steps of the process of grinding, and the nature of the invention also resides in the derivation and use of certain mathematical proofs which aid in the correlation of some of my newly discovered facts and in the determination of others which may otherwise be unknown. More particularly, an observation of a grinding mechanism or mill, shows that it discharges some solids which are ground finer than is desired; some sized with the limits that are desired; and others that are larger than desired.

Of these, the properly sized solids are removed for subsequent treatment, while the oversize solids must be returned to the grinding mill for further grinding. They represent incompletely ground solids on which more work must be done by the mill. On the other hand, the undersize represent solids on which too much work has been done for they have been needlessly and uselessly ground finer than is necessary. Therefore, one feature of this invention is to control the proportion of these different sized particles produced by the grinding mill. That is, it is desired to control the surface producing force of the mill to produce the desired particle size and to prevent the production of overground sizes. This I accomplish by circulating thru the grinding mill, along with the oversize, a certain sized slime which is beneficial to grinding and to classifying efficiency.

-A further observation discloses that there are certain critical conditions under which the grinding media of a mill will operate efliciently. If a mill be crowded with clean sand, the friction and compacting effect is sufficient to markedly decrease the mobility of its rods or balls. The mass moves and acts much the same as when a pudding is dumped out of a pan. The grinding media merely tends to move in a mass of sand which resists mobility. Such grinding as is offected, is largely produced by the dropping impact, which is the least valuable factor involved in grinding. It has been found necessary to cause an adheslveness between the grains to be ground and the balls, rodsor grinding media of the grinding mill. If the balls can be caused to become and maintained coated with sand grains, then when two or more balls contact, a. distinct crushing of the coating grains takes place. Therefore, my aim is to set up a plasticity of the mass to be ground whereby particles thereof adhere to the balls or other grinding media. This I accomplish by circulating thru the grinding mill with the material to be ground, a slime of selected particles of a 'size conducive to this condition.

If a mill be crowded with clean sand, the balls lie in an immobile mass of sand, as above explained, and the mass of sand and balls tends to rotate as a unitary body. As opposed to this, my aim is to produce a free movement of the balls with respect to the sand. The character of the ball action is required to be different. The balls should cascade with respect to the sand. The entire mass of balls and sand are desired to be mobile with respect to each other. Each ball should rotate *and spin in its contact with other balls or liners. At each contact the coating sands keep a constant feed between the balls. Therefore, my aim is-to produce in the mill a condition of lubricity between the sands and the balls. This I accomplish by maintaining in the mill a selected fraction of slime including certain sized solids favorable for obtaining the desired result.

With those explanations, it can be seen that this invention resides in setting up and maintaining in the grinding device, a condition of plastolubricity which is brought about in a simple manner by feeding to the grinding mill, slime having certain characteristics, the chief of which is i that it shall contain as many as possible of colloidal-like particles. In theory these particles should be of the diameter of a micron or less, but in practice I aim to useparticles having a diameter up to and includihg that of a suspensoid which sometimes is called a suspension. An ideal concurrent condition would be to have the slime substantially free of particles between a micron and oversize, but this is not possible in commercial operating conditibns, so I approach this condition as nearly as possible under operating conditions. This slime may be added to the mill as a separate substance, such as bentoniteL However, in tests with closed circuit grinding, it has been shown that in grinding practically every substance capable of such treatment, the grinding ticles. So under all normal conditions, no extraneous material is likely to be needed toget the required slime conditions, at least after they have been first built up within the grinding circuit.

But one feature of the required slime condition is that there be added to the mill, a minimum of unfavorably sized particles or grains between those of a micron in size and those which are of the desired mesh for which grinding is being done. For example, suppose we are grinding to et a 48 mesh product. There will be in the mill solids ranging from molecular size to those up to suspensoidic size. Let us call that range A. Then there will be other solids falling in a range which will be called B including sizes from suspensoidic to say 65 mesh. There will also be solids of +65-48 mesh, which will be called C. And those larger than 48 mesh, or oversize, will be called D. It is those in the range B that this invention teaches should not be recirculated to the mill, for they are already small enough so the mill should not be burdened with them. It is the quantity thereof which should be kept down to a minimum. It is the grains in the range A which I refer to as the favorable fraction of slime that I want to have continually present in the mill. Those in the range C constitute the finished mill product, while those in the range D must be recirculated to the mill for further grinding.

This invention therefore, also contemplates the separating or selecting out from the mill discharge, solids in various size ranges whereby those in range C are removed from .the circuit and are ready for some type of subsequent treatment; those in range B are removed from the circuit and join those in range C as finished product, while those in ranges D and a necessary portion of those in range A are recirculated through the mill along with the new mill feed. So a most important point of this invention lies in recirculating to the system, or through the mill, slime having colloidal types of solids within size limits up to a suspensoid as Well as oversize, while eliminating as far as possible from the grinding devices are contemplated thereby. Ac-

cordingly, that material on which the ball mill operatesis referred to herein interchangeably as sand, grains, particles or solids. Associated with a ball mill in closed circuit grinding systems are classifiers but when classifiers are referred to herein, any mechanism which has a classifying, stratifying, or sorting action, is meant. While any type of classifier, such as the conetype, may be used, it is preferred to use a raking type so that type will be referred to. Herein the discharge end of a classifier is considered where the raked product or sand is discharged to emergence and the overflow is the other end of the classifier or where the slime pulp overflows. The deep end is called the overflow and the shallow end the discharge. Whereas, slime is referred to herein, itis intended to cover any such material as colloids, clay, chemicals, oil, and so forth, so long as they produce in the mill, the desired conoperation automatically produces slime, some portion of which, contains the desired size of pardition of plasto-lubricity when introduced at any point in the circuit. And it is understood that the slime either of one kind or of several kinds may be added to the circuit at any point, if for any reason there is not sufficient produced in grinding the material being treated.

In the same manner, the process of this invention has wide applicability so it is intended that its use shall be protected hereby in any industry where it is desired to grind while wet. Specific materials which may be so treated are ores, rock products, phosphates, fertilizers, cement, furnace products, coal, chemicals, clays, sands and, pigments. While the invention has been developed for use in closed circuit grinding, it may find useful application to grinding in other than closed circuit but in order to show the manher in which the invention may be carried out, a closed circuit operation will be described. To that end, I have accompanied this specification with drawings for illustrative purposes. In the drawings, Figure 1 shows diagrammatically the manner in which the mill discharge is subdivided and treated, Figure 2 shows a flowsheet indicative of a workable arrangement of equipment wherein this invention may be carried out. Figure 3 shows diagrammatically how a balanced condition is set up in the secondary classifier Y. Figures 4 to 6 show testtubes in which various samples have been taken. Figure 7 is a diagram to be used in connection with the mathematics set forth herein.

The advantage of a process of controlled size grinding has long been recognized. In the grinding or crushing, of an ore for example, it is necessary to crush to that degree of fineness which will free the valuableminerals from each other as well as from the valueless or gangue minerals. In the subsequent separation of these parts it has ,been found undesirable to have any large portion of them too finely ground because the recoveries will thereby be adversely affected and energies will be wasted. For example, selective flotation works best both for recovery and grade of concentrate, when operating upon a mineral of reasonably uniform size if that mineral has been crushed just fine enough to liberate or free the valuable constituents from a worthless gangue and from one another. Any grinding in excess of this will show poorer recovery and poorer grade, and at the same time, consume extra energy in grinding, thereby increasing the grinding costs. If the mineral, after grinding, is to be treated in a hydrometallurgical way or is to be roasted, it is essential that there should be no overg'rinding which would result, for hydrometallurgical uses, in slower settling or thickening; slower or more inefiicient filtering and washing; and in roasting operations, in greater dust losses.

Total handling costs are thereby decreased by the method of grinding control of this invention in both capital investments and operating costs.

It has long been accepted that slime should be kept from circulating thru a closed circuited mill. This invention upsets that theory. This invention teaches that the presence of slime in a mill is essential for highest mill efficiency. But more particularly, I have found that not all slime is helpful but instead only a favorable fraction thereof is beneficial. That favorable fraction is not one of quantity so much as it is one of particle size. The ideal or theoretical size of the particles making up the favorable fraction of slime is that they be a micron or less. In practice, this fine line of selection is very diflicult to make. If by chance the ideal range of sizes cannot be obtained in practice, then the next best range of size of particles lies from the molecular up through the colloidal and that borderline type of material lying between the colloidal and suspensoidal, which may be called the quasi-colloidal; But in commercial practice, the favorable fraction is a concentrate of the finest sizes of the material in process.

If the slime discharging from the mill be separated or concentrated out from the entire mill discharge and then exposed to a further selective action whereby there is collected, as nearly as can be done under commercial conditions, a group ofslime solids containing a maximum of colloidal and a minimum of larger size particles,

this collected and selected portion of the slime constitutes that fraction of the slime which is favorable to grinding. It is this favorable fraction of slime which this invention proposes should be circulated thru the mill and classifier with its circulating load. The unfavorable fraction of slime constitutes that portion of slime rejected. in the selection just referred to and it will contain a majority of particles larger than a colloid.

While the essence of the invention therefore is this recirculation of a favorable fraction of slime, the manner in which this favorable fraction is separated or isolated, and selectedor applied, is

also an important part of the invention.

In a simple application, then, the invention may be carried out in a grinding mill circuited with a classifier. The mill discharge is fed to a primary classifier. This classifier separates out the oversize which is returned to the mill and the overflow from the classifier contains the desired fines plus slime. This overflow is then exposed to further separation or selection such as by classification or hydroseparation where the desirable fines plus the fraction'of slime unfavorable to grinding are discharged at the sand end of the classifier and the fraction of slime favorable to grinding is recovered from the slime end of the classifier.

This favorable fraction preferably is returned to the mill for recirculation therethrough'. The discharge from the secondary classification or separation contains the other or unfavorable fraction of the slime plus the end product sands and some portion of the favorable fraction of slimes, the mixture of which form the finished product of the circuit. This proportioning is exemplified in Fig. 1- As will be shown later, my invention provides simple means for controlling a proportion of the favorable slimes in the system, which are eliminated from the system with the finished material, and of keeping the amount of slimes so eliminated equal to the amount of slimes produced in the mill.

But in practice, it is not as easy to do as might appear. Certain conditions must be caused to exist in the classifying stages in order to obtain the precise selectivity required.

Referring to Fig. 2, and to the fiowsheet shown therein, the new mill feed enters the ball mill and so does the circulating load. The mill discharge is composed, as usual, of oversize, fines and slime. The discharge goes to a primary classification stage or classifier X. This classifier makes two separations. The oversize sands are separated and discharged by the classifier from whence they are returned to the mill as circulating load. The separated overflow from this classifier contains the fines plus the slime produced by grinding in the mill and what slime is' circulating in the system.

This overflow from the classifier X, composed of fines and slime, is then fed to a secondary classification or other separation stage such as a classifier Y. In this classifier, the two portions or fractions of slime are selectively separated, the unfavorable fraction of which with the finished fines are discharged and are ready for use. The favorable fraction of slime may preferably be overfiowed from classifier Y into a compartment or reservoir Z from whence it goes to recirculation to the ball mill or classifier, or to both. The unfavorable fraction, composed of the larger sized slime solids are thus discarded or eliminated from the circuit and thus are precluded from getting back into the mill. They are ejected or rejected from the system or circuit by being mixed with the finished fines and exit with them.

The circulation thru the mill of this favorable fraction of slime produces thevarious desirable conditions in the ball mill hereinbefore referred to and which I term plasto-lubricity. That is, the individual grains in the mill adhere to and coat the balls and yet the balls and the mass of sands are mobile and cascade freely independently of each other.

Referring now to the classifier X, greater selectivity is required than is ordinary. This can be accomplished in various ways. I propose now to make certain modifications of the classifier. One constitutes increasing the density of the classifier bath in order to refine the selectivity or efficiency of classification of the classifier. This increase of density of classifier bath may be done in any suitable manner, altho my preferred plan is to return to the classifier bath some of the favorable fraction of slime. The favorable fraction of slime tends to have a constant density. In one ore treated, the density of the recirculated fraction of slime was maintained fairly constantly at 1.23, but of course this can be varied to meet the requirements of different ores or materials being ground or to meet different conditions in grinding the same material.

The reason for this increased density of classifier bath is to endeavor to preclude the passing to discharge of any undesirable larger particles or unfavorable fraction of slime. The discharge from this classifier is oversize which returns to the mill for regrinding, so if this discharge should include any appreciable quantity of the unfavorable fraction of slime or of critical sized fines the aim of this invention tends to be defeated, for one feature of the invention is the keeping out of the mill those larger sizes of slime solids which I have found to prevent efiicient grinding.

The increased density of the classifier bath also serves as a determinant for the critical mesh or sized fines which are usually uncertain just which way to go-to discharge or to overfiow. The increased density determines for them that they are to go to overflow and away from the discharge end of the classifier. So by increasing the density of the classifier bath I have found that improved selectivity of the classifieris obtained and it can be depended upon that most of the critical mesh sizes will be kept out of the oversize sand discharge.

I may also control the selectivity of this classifier by supplying to the classifier a certain volume of the favorable fraction of slime, and also by increasing'the velocity of overfiow over the end weir. This increased volume of slime is so modified is that there are dispersed the critical sizes of fines which would otherwisev feed, and the increased volume of overflow causes a dispersion of the critical mesh sizes in the classifier which makes for cleaner classification. Thus I contend that in a classifier, I can improvethe selectivity of classification by increasing the density of the classifying medium coupled with an increase of volume of fluid fed thereto, and an increase of velocity of overflow. In this particular instance, an .increase of volume of fluid fed the classifier also produces increased velocity of overflow, but I can conceive of circumstances where other means than increased volume of fluid feed is necessary to increase the .velocity of overflow. That is, I propose to improve classification or selectivity of classifier X by doing three things, namely, circulating slimy fluid to the classifier to increase the volume of fluid passing therethrough; increasing the velocity of overflow over the end Weir which may also be a submerged weir; and increasing the density of the classifying medium. While increasing the density of the classifying medium, I may, by recirculating slime, increase the dilution of the classifier overflow, since the slime recirculated will normally have a higher dilution than the classifier overflow, as they form a selected portion of the overflow solids together with a major portion of the water and have a density lower than the primary classifier overflow. At the same time the ratio of slimes to water in the primary classifier overflow is maintained relatively the same if not increased.

The characteristics of the pump'in terms of density, viscosity, volume and dispersion are now juxtaposed for the sake of convenience:

(1) Density of the pump is the concentration of total solids.

(2) Viscosity of the pump depends mostly upon the concentration therein of the finest solids or slimes.

(3) Volume of the pump determines the velocity of flow through the tank between the point of feed and the overflow, dispersion of particles of course being also a function of volume. Another modification of this classifier X may be the elimination of fresh wash water, as such, on the upwardly raked sands because this would dilute the-bath and prevent the building up of its gravity or density, which is much desired.

Apparently, what happens when the classifier normally be crowded close to one another, and this condition would tend to prevent the coarser particles from sinking and the finer particles from rising to the overflow. An increase of the volume of classifying fluid in which these critical sizes are suspended achieves their dispersion or causes them to separate one from the other, and this assures their proper classification and subsequent passing to overflow. The density, the volume of fluid fed, and the velocity of overflow must be carefully controlled for the particular material being treated so that the density will be high enough to cause-the assured overflow of the critical sizes of fines but not high enough to carry oversize to overflow. But once the proper control is found for the material being treated, little attention need be given it further, for once set, the conditions of the bath remain fairly constant.

The overflow from classifier X, which is a mixture of fines and slime, 'then goes to further classification or selection. This further selection may be carried out in a number of ways but by 1 classifier. .'up in that classifier of a condition of automatic way of illustration, let us assume this to take place in a secondary classifier Y. Here classification takes place but the operation of this classifier is controlled so that the finished fines are discharged and so are the unfavorable fraction of slime solids. That is, this secondary classifier Y, or any equivalent thereof which may be used, is depended upon to select out and to overflow the favorable fraction of the slime solids from the unfavorable fraction and the fines. The dependability of this separation or selection is important to the successful carrying out of this invention. So to be doubly sure that the selection of the favorable fraction of slime solids will be carefully accomplished, wash water is applied to the secondary classifier Y, as indicated on the flowsh'eet shown in Fig. 2. This sheet of wash water serves as a screen for holding back away from the discharge end, the favorable fraction of smaller slime solids. This assures their going out the classifier overflow and not becoming lost to the circuit by going out the discharge end along with the fines and the unfavorable larger slime solids. This classification and wash water also assure the unfavorable slime solids settling out whereby they are prevented from getting into the overflow and thus back into the circuit where they are so harmful. This wash water also has the added function of restoring to the circuit, the water which passes from it'through the secondary classifier discharge with the finished product.

In one mill circuit the finished product from the secondary classifier hada dilution of 50% or 1:1. Thus the end product is unusually dry. This precise amount of water has to be returned to the circuit but as the amount needed is fairly constant, it is a simple matter to set the new water being fed to the circuit, for it is merely correlated with the new mill feed on a 1:1 basis. If the finished product is to have a different dilution, the ratio of new water to new mill feed should be varied accordingly.

Reference has been made to the constant density of the recirculated slime and to the constant dilution of the sand discharge, both flowing from the secondary classification stage or This is accomplished by the setting balance. We have at the overflow end of the classifier a zone of greater density than in the zone at the discharge end. This is because of the wash water, which being applied substantially at the zone of emergence of the classifier blades, keeps the,flnes and slime washed back away from that zone, whereupon the zone of emergence of the blades becomes a zone of greater dilution. As the wash water washes back the slime and fines into the overflow end of the classifier, their concentration at that end sets up therein and thereat, a zone of increased density which is materially greater than the density of the zone of dilution or rake-blade emergence.

In Fig. 3 classifier Y is shown more or less diagrammatically in which ll indicates the inclined bottom or deck of the classifier, l2 the overflow end, l3 the submerged weir, l4 the discharge end, and I5 the raking blades or other mechanism. The submerged weir l3 divides the'classifier tank into two compartments or sections, .in one of which the raking blades I5 operate. ifier overflows into the other compartment or section Z, in which the favorable fraction of slime flows and accumulates, and from which The classi emergence of the raking blades l5 which is also the zone of dilution, while [9 is the zone of in creased density. The wash water is added at 20. The normal horizontal liquid level for the classifier is indicated in full lines at 2|, while the actual liquid level line assumed by the classifier bath under balanced conditions is indicated by the dotted line 22. As the density in the dense zone I9 is built up, it naturally takes more dilute liquid in the dilute zone [8 to balance it. In order to get enough of the dilute mixture in the zone I3 to balance the dense zone, the liquid level of the dilute zone rises and produces a super-elevation.- In effect, it is as if the classifier were a U-tube with the denser zone in one arm of the tube and the dilute zone in the other. Under these conditions, it can be visualized that the liquid inthe dilute arm will be higher than the dense liquid in the other arm, if the liquid in the two arms is to be in balance. This is exactly what happens in this classifier, -so the liquid level in the dilute zone I8 is automatically maintained higher than in the dense zone IQ, for they thus balance each other. Under operating conditions, of course, with the incoming wash water there is a tendency for the liquid level in the dilute zone l8 to build up higher than is necessary to balance the dense zone I9, and when .thishappens, the excess water passes out of the discharge end M of the classifier along with the raked product, because the classifier is so levelled that the discharge end intersects the plane of the balanced liquid level 22. But at the same time, this slight rising of the liquid level in the dilute zone reacts to raise somewhat the liquid level in the dense zone, and this slight rising thereof, causes an overflow over the weir l3 of enough slime to bring the liquid levels back into balance. Thus there goes on a constant building up of the level in the dilute zone which acts partly to discharge some of the wash water with the raked product, and partly to raise the liquid level in the dense zone whereby there is produced an overflow of the slime. However, one zone balances the other, so the two zones quickly, constantly and automaticallyreturn to equilibrium. It is this automatically balancing condition within this classifier that causes the withdrawn slime to be of constant density. As has been stated, the slime withdrawn from this classifier constitutes the favorable fraction thereof which is circulated through the circuit, while the emerged discharge from this classifier constitutes the unfavorable fraction of larger slime solids and the fines, both of which form the finished productof the circuit. The selected favorable fraction of slime overflowing the weir of classifier Y is collected in some suitable container such as compartment Z from whence it is fed back to the mill or other desirable place in the circuit. The liquid level in the compartment Z is the same as that in the back end of the classifier adjacent toor abutting compartment Z and the weir I3 is submerged. This is necessary if, for

any reason, an increased amount of liquid is withdrawn from compartment Z, the entire liquid level of the'classifier drops momentarily and then all of the wash water is promptly retained until the liquid volume is restored and balance is again attained. In my preferred form of arrangement, some of the favorable fraction goes to the mill and some to the classifier X, as has been explained above.

Fig. 4 represents a test tube in which has been caught a sample of theoverfiow from the first or primary classification stage, in which the component parts are shown as they appear after a brief period elapses for sedimentation. 33 indicates a colloidal suspension, 3| a settled colloidal like mass, 32 the finest sizes of granular particles and 33 the larger fines up to the desired mesh of the grind.

colloidal suspension. But some ores, or some I mill circuits may contain a fiocculating agent. If so, the colloids normally in suspension in this portion are flocculated and settle into the settled colloid like portion 3|. Itis immaterial to the functioning of this invention whether the colloids or slime are fiocculated or not. Such changes as are efiected by flocculation are met by other adjustments of the system under control. The portions 30 and 3| form the ideal or theoretical favorable fraction that is preferred for recirculation in the mill circuit, but in practice, this fine a line of division or selection cannot be made. Some granular fines usually get into the favorable fraction as shown at 32, but it is important that as many as possible of the granular fines 32 be rejected from recirculation.

some comparative tests obtained with slime andwithout are here given:

Test N0. 59 60 53d Slime returned. No Yes Tons of original feed.. 12. 7 13. 6 Circulating load 13. 6 3. 8

Finished product Mesh Cum. Cum. Cum

By knowing the laws governing this process, it is possible to materially increase the tonnage ground in a given mill, and because of the total advantages to be gained, to materially decrease Where T=tons per day of new mill feed or ton- Then the number of final cubes is r and thefinal surface is The new surface is:

1 Fa --6E =6E*( r-1)" If a ton of ore originally contains A grains of mean edge E and this is ground until the mean ratio of reduction is r, then the new surface created is 6AE (r1).

Let the surface creating force be P, measuredin terms of surface, and let the number of tons of ore ground be'T, then is another constant. (7) Hence (T- 1)(1+C)-i$ a constant, and

From Equation (5) (11) T=K (1+C'). Let K =K1 a constant, and (1+C')=R, where R is the total feed ratio to grinder:

(12) Then, T=K1R.

Equation (8) is practical because T(1+C)= total ball mill feed is known. Equation (10) may prove of theoretical interest.

Equation (7) must be understood to mean that c at any given size classifier overflow. such as %5 mesh, for instance, 1' is a variable depending on the circulating load carried in grinding the ore to 48 mesh. If the circulating load is low, T must be larger and this can only mean that there must be a larger amount of overground material in the finished'product. This should be indicated by a higher percentage of -200 mesh material in the product from a grind with a low circulating load as compared. with the same grind with a' high circulating load. 1

In other words, the equation indicates that the tonnage ground to any specified mesh in a given ball mill with a feed of constant size should rise with an increasing circulating load and that this high tonnage is obtainedby decreasing the amount of overgrind.

Returning to the equation it will be recalled that the constant value of'K (the mill coefiicient) depends on the factor -oo' +0) If this factor varies, K necessarily must vary. If it were a constant, K necessarily would be a.constant. Recalling that]? is the amount of surface created in a ball mill, then anything that happens in a mill which changes the efliciency with which the power is converted into surface,

affects P, and consequently affects K.

With increasing circulating loads, especially if the classifier efiiciency is reasonable, slimes are removed and prevented from entering the mill but no matter how much natural slime an ore has conditions do arise whereby insuflicient slime is in the ball mill circuit. P necessarily must be adversely afiected and consequently the mill con-' stant will likely decline in value. 7 On the other hand, this is not at all an argument in favor of low classifier efiiciency. The

return of slime to the mill is favorable, but if much critical mesh sand is returned that is sumciently'ground, over-grinding will be pro moted and this is decidedly unfavorable.

The point it is desired to stress here is that plasto-lubricity is probably the greatest factor affecting the element P and consequently the coefficient K. It K is to be maintained reasonably constant, especially with increase in total mill feed, care must be taken to maintain adequate slime in the mill itself.

In simple languagathe above formulae, which have been supported by evidence, show that under any particular set of conditions, the tonnage ground is proportional to the square root of the load circulated According to this formula, the circulating load ratio could be increased indefinitely and theoretically an increase in tonnage ground would be obtained, such increase in tonnage being proporthe coarsest particles, or in other words, the ratio of reduction tends to approach one.

Some other factors which affect the element P and the coefiicient K, are generally incident to operation, but are worthy of note. They are: (1) Change in weight of ball charge; (2) change in ball assortment; (3) change in moisture content; (4) change in ball mill speed; (5) change in size of ore feed; (6) change in degree of agitation of ball charge occasioned by lifter bars; (7) change in design of liners; (8) change in mill diameter due to liner wear; (9) change in the specific, gravity of the ore, whereby less surface is required to reduce a given weight of ore through a given mesh; (10) change in amount of finished material in the mill feed. This, of course, only apparently affects the value of K; and (11) change in the plasto-lubricity. The latter is a repetition but may as well be in the list.

Operation.--In order to make the operation of this invention clear, let .us assume a new closed circuit mill has been constructed and is ready to 'be started up. The operator will load up his mill and start it going along with the classifiers X and Y. However, his first object will be to build up in the circuit an ample supply of favorable fraction of slime. This could be done either by adding or introducing slime with appropriate characteristics to the circuit, or, by directing all of the water being added to the circuit to flow thereto through the wash water pipe 20.

A' simple test as to whether or not there is an ample supply of slime in the mill is to catch a sample of the mill discharge in a bucket. After a moment or two, the bucket is upset. If the sands stick in the inverted bucket. there is insufficient slime in ,the mill, but if the sands pour out of the bucket, there is suflicient slime present. This test is as good as any laboratory test for practical purposes. In tests wherein an ideal colloidal slime was added to the mill in the form of bentonite, it was found that if the bentonite present constituted 2% of the circulating load in the circuit, it was sufficient. But bentonite is an ideal favorable fraction of slime for it is not diluted with any granular fines. So in proportion as the colloidal slime of the favorable iraction is diluted with granular fines. so that much more must be used. If the favorable fraction being used is composed of colloidal slime equal in effect to bentonite and 50% granular fines, 4% must be used based upon the circulating load. and so on.

As soon as an inspection tells the operator that there is enough favorable fiaction of slime in the circuit to give him his desired mill efliciencies. and he observes that his classifier Y has reached a balanced condition, then he cuts down on the amount of wash water being fed to classifier Y and by-passes the remainder of the new water required by the circuit (to make up for the ater losses going out with the raked end product) in the whole system for otherwise all of the conditionsappear to be automatically adjustable. He watches .his feed water to make sure it is just enough to keep the classifier Y in operating balance. If he does this, the density of the discharge product from classifier Y contains the fines plus the undesirable or unfavorable fraction of slime solids and its dilution is constant at whatever is wanted, such as 1: 1. This also gives him the favorable fraction of slime withdrawable favorable sized slime or fines and also assures him of aconstant density of the favorable fraction of slime. This favorable fraction of slime is circulated to the mill, to the classifier X. or to both, as may be found to give the best milling conditions wherein the grains stick to and coat the grinding media, and the grinding media and the mass being ground are freely mobile with respect to each :other. 'In other words. to produce the conditions in the mill of what I have termed herein plasto-lubricity.

He also checks the overflow from classifier X to make sure that conditions therein are correct. That is, to see that its bath is dense enough to assure proper selectivity thereof, as well as the volume of infiowing or infeeding slime thereto and also the velocity of overfiow, for it is imperative that a maximum of critical mesh sizes and fines are kept in the overflow. If this classifier is not classifying to the degree required, he brings it to proper functioning usually by feeding a greater quantity of the favorable fraction of slime from the classifier Y overflow (or compartment Z) until conditions are right. If too much slime is being discharged with the r ked oversize in classifier X, he will cut down on the quantity of favorable fraction of slime being fed.

to the classifier until the slime and fines are being properly overfiowedby this class fier X. And that, along with any desirable variation in rake speed, is about all the operator needs to do for once the circuit is function ng well. it tends to continue to do so with very little attent on. The circulating load seems to take care of itself and automatically adjusts itself to any variation in the ore being ground.

One word more about the favorable fraction of slime. It is difficult to clearly express what is meant by this term. It is difficult because the theoretical or ideal conditions are that the favorable fraction should contain a preponderance in number of particles of a micron or less in s ze (30 and 3| in Fig. 4) with an absolute minimum of larger sized solids. But this degree of selection is extremely 'difiicult to obtain under actual commercial scale of operation. So, we have to do the best we can.. That-best seems to be that we select out or concentrate from the material in process a favorable slime bearing fraction thereof which comprises the maximum commercially possible of particles whose size ranges from the molecular up through the collo daland qu sior semi-colloidal sizes to those which approach the critical mesh sizes for the system. (30, 3|, and 32 in Fig. 5). In practice, the favorable fraction is that fraction of the ore smaller in maximum particle than the second finer mesh Tyler screen. For example, at 48 mesh separation the favorable fraction is all l00,mesh. At mesh separation, the favorable fraction is all -200 mesh. At 35 mesh separation, the favorable fraction is all -65 mesh. But it is desired that this importance attached to the particle 'size in the favorable fraction does not blind the reader to the ultimate aim of the use of that favorable fraction which is to materially increase mill efficiencies by creating therein plasticity which is that condition of adhesiveness created in the mill which causes the grains to stickto or coat the grinding media, and by creating therein lubricity from the classifier containing a minimum of unmill for regrinding. ,Only one \more observation will be made about the favorable fraction of slime and that ispthat so far I havefound no upper limitfor the amount of the favorable fraction that can be circulated through the mill or the system, and also that the more favorable fraction that is fed to the mill, the less slime the mill seems tocreate out from the material being ground.

The advantages to be secured from carrying out this invention are many and include materially increasing the effective capacity of the mill. The cost of grinding is reduced because less power per ton ground is used. .Higher circulating loads can be used in the grinding circuit. A better final product is obtained because it is'granular rather than slimy. If. ores are being ground, there is decided improvement in the metallurgy for there is an increase in the recovery of valuable material from the ore due to the absence of excessively fine material; better thickening is possible and so-is better filtration. Thus the benefits of this grinding process progressively increase as the ground product is successively exposed to metallurgical treatment. Less water is used in-the circuit so a drier end product is obtained. As the ore is drier, the capacity of equipment, for example, a flotation cell, to which the ore is to be exposed is materially increased. Further, the tailings are more granular and are easier impounded which contributes to a lessening of stream pollution. Corresponding savings and advantages are also experienced when the process is used in other than metallurgical industries, such as in cement making.

I claim:

1. The process of closed circuit grinding wherein a mill is used comprising adding to the mill feed slime having therein a preponderance in number of solids whose size is less than that of a suspensoid, and rejecting from the circuit slimes having therein larger sized solids.

2:- The process for use in a grinding mill cirtreating said slime to selective classification to separate therefrom a fraction thereof in which there is a preponderance in number of solids of a size of a quasi-colloid or smaller which is favorable to grinding efliciency, and then circulating such fraction and oversize material through the mill.

3. The process of closed circuit grinding in which a mill is used comprising treating the mill discharge which includes slime to selective classification to separate said discharge into oversize on one hand and on the other slimes and fines, classifying the latter into two groups of material, one group composed of sub-suspensoidal sized slimesolids, with the other group composed of suspensoidal and larger sized slime solids and fines, rejecting from the circuit the latter group, and circulating through the circuit material of the former group.

'4. The process for use in a, grinding mill circuit comprising separating the slime from the mill discharge which includes slime, selecting out from the slime a fraction thereof favorable to grinding a preponderance of whose particles are quasicolloidal or smaller, and circulating a quantity of said favorable fraction of slime in the circuit with said quantity being at least two percent of the total feed to the mill.

5. The process for usein a grinding mill circuit comprising separating the slime from the mill discharge which includes slime, selecting out from the slime a fraction thereof favorable to grinding the preponderance of whose particles are suspensoidal or smaller in size, and circulating said fraction of slime through the mill.

- 6. The process for use in a grinding mill circuit in the mill discharge of which there is slime comprising dischargingin a c'lassification stage the oversize from the mill discharge, circulating said oversize to the mill, subjecting the remainder of the mill discharge to a further classification stage to separate therefrom slime in which there is a preponderance in number of particles substantially colloidal in size, and circulating such separated slime in the mill circuit.

7. The process of grinding in a. closed circuited mill in the mill discharge of which there is slime comprising selecting out from the mill discharge a quantity of the slime therein, and further selecting out from said slime a portion which if sampled and allowed to settle in a test-tube will show a. substantial part of the sample to contain colloids, and then returning said portion of slime to the mill circuit.

8. The process of grinding in a closed circuited vmill in the mill discharge of which there is slime comprising selecting out from the mill discharge by two steps of selection a fraction of the slime in the discharge which is favorable to grinding a preponderance of whose particles are suspensoidal or smaller in size, and returning said selected fraction of slime to the mill circuit, while rejecting'from the mill circuit the remainder of the slime and fines in the mill discharge.

9. The process for use in a grinding-mill circuit in the mill discharge of which there is slime comprising separating the slime from the mill discharge in a classification stage in which the classifying medium has an abnormal density, and returning a portion of said slime to the circuit which is at least two percent of the mill feed.

10. The process according to claim 9 in which said slime portion is returned to the classification stage.

11. The process of grinding in a closed circuited mill in the mill discharge of which there is slime comprising subjecting the mill discharge to a primary stage of classification to select therefrom the slime and fines, subjecting said slime and fines to a secondary stage of classification to select a desired fraction of slime from the slime and fines, and maintaining a relatively quiescent body of liquid having slimes therein in direct hydraulic communication with the secondary classification stage through a common upper strata as defined by a submerged weir whereby the selected desirable fraction 'of slime is derived by Way of said quiescent body at a'point substantially spaced from the point of hydraulic communication.

12. The process of grinding in a closed circuited mill in the mill discharge of which there is slime along with oversize, which comprises exposing the mill-discharge to continuous selective classification to separate therefrom oversize tionhaving a constant density continuously reon the one hand and on the other fines and slime turning such selected uniformly dense slime to continuously selecting by classification a. portion said circuit together with oversize from the classiof the slime having therein a. majority of particles flcation step, and rejecting hues and slimes other 5 of a, size up to and including that 0! a susthan the last-selected slimes from the circuit. R 5

pensoid with said continuously selected slime por- ARTEUR JOHN WEINIG.

. CERTIFICATE or CORRECTION. I Patent No. 2 ,1ol 709. January i 19 ARTHUR JOHN WEINIG.

It is hereby certified that the above numbered patent was erroneously issued to "The .Mine 8c Smelter Supply (30., The Stearns-Roger Manufacturing Con, and The Dorr Company, Inc."-, as assignees, whereas said patent should have been issued to The Stearns-Roge'r Manufacturing- Company, of Denver, Colorado, and The Dorr Company, Inc., of New York N. Y., assignees by direct andmesne assignments, as shown by the record of assignments in this office; and that the said LettersPatent shouldbe readwith this correction therein that the same may conform vto the record of the case in the Pat nt Office.

Signed and sealed this 26th day of DecemberpA'. D. 1959.

. Henry Van Arsdale o (Seal) Acting Commissioner of Patents.

tive classification to separate therefrom oversize tionhaving a constant density continuously reon the one hand and on the other fines and slime turning such selected uniformly dense slime to continuously selecting by classification a. portion said circuit together with oversize from the classiof the slime having therein a. majority of particles flcation step, and rejecting hues and slimes other 5 of a, size up to and including that 0! a susthan the last-selected slimes from the circuit. R 5

pensoid with said continuously selected slime por- ARTEUR JOHN WEINIG.

. CERTIFICATE or CORRECTION. I Patent No. 2 ,1ol 709. January i 19 ARTHUR JOHN WEINIG.

It is hereby certified that the above numbered patent was erroneously issued to "The .Mine 8c Smelter Supply (30., The Stearns-Roger Manufacturing Con, and The Dorr Company, Inc."-, as assignees, whereas said patent should have been issued to The Stearns-Roge'r Manufacturing- Company, of Denver, Colorado, and The Dorr Company, Inc., of New York N. Y., assignees by direct andmesne assignments, as shown by the record of assignments in this office; and that the said LettersPatent shouldbe readwith this correction therein that the same may conform vto the record of the case in the Pat nt Office.

Signed and sealed this 26th day of DecemberpA'. D. 1959.

. Henry Van Arsdale o (Seal) Acting Commissioner of Patents.

Images