Another modification of the White furnace has been made by J. M. Thompson.

This furnace differs from the Howell-White in several particulars. The cylinder is of one diameter throughout its length, and is lined throughout with tiles, thus lessening the loss of heat by radiation, which must be considerable in the unlined portion of the Howell-White furnace. The tiles are specially molded for the purpose, and are so formed that the projections and recesses, designed to lift the ore, are of a rounded form as to their transverse section. In some cases the recesses are obstructed at intervals by transverse ridges. The intention of this is to prevent the ore sliding along in the inclined cylinder, in consequence of its low angle of stability at an early stage of the roasting. In some ores this angle becomes almost nil, so that the powder resembles a fluid in its action.

The further retention and equalization of heat in the cylinder, is promoted by means of a layer of non-conducting material, such as paris plaster, or asbestos, between the lining, of bricks or tiles, and the iron shell. The latter, and the supporting rollers with their bearings, are thus, in a measure, protected from the injurious effects of unequal or excessive heat, by being kept comparatively cool.

Thompson's Improved Howell White Furnace.

Thompson's Improved Howell White Furnace.

The furnace is arranged in either of two ways. In the one, the whole of the ore enters the cylinder at the higher and cooler end. In this case the lighter dusty portion which is carried back by the draft, is roasted by the flames from an auxiliary fire between the end of the cylinder and the dust chambers, the latter being constructed of masonry. In the other, the dust is separated from the coarser portion of the ore before entering the cylinder. The coarse portion then enters at the upper end, while the dust enters at the lower end near the fire. The main fire is thus made to serve for the roasting of both portions of the ore, and the auxiliary fire is dispensed with. It is also claimed by the inventor, with considerable show of reason, that the heat and the chlorine, developed by the roasting of the dust, mingled as it is with salt, instead of being lost, as when an auxiliary fire is relied on, aids materially in chloridizing the coarser portion of the ore.

A portion of the dust, impelled by the draft, traverses the entire length of the cylinder, and entering the flue is arrested in its flight by the ordinary means of dust chambers, which, in this case, are constructed of iron instead of masonry; hence they are light, portable, and cheap.

The length of time occupied by the passage of the coarser portion of the ore though the cylinder, is regulated mainly by the inclination of the latter, which, as in the White furnace, is adjustable, though, as the writer is informed, in a different manner. Mr. Thompson prefers to regulate the period of transit by varying the angle of inclination, rather than the speed of rotation, of the cylinder, arguing that, as it is stated by eminent metallurgists that the loss of silver by volatilization is, other things being equal, proportioned to the length of time during which the ore is exposed to the heat, it follows that the more quickly the ore is roasted the less loss it will sustain. But, he continues, the greater or less rapidity of the roasting depends on the greater or less exposure of the particles of ore to, first oxidizing, then chloridiz-ing influences, or to both simultaneously; hence the more continuously the ore is showered through the heated air and gases pervading the interior of the furnace, the more quickly will the roasting be effected, and consequently less loss will be caused by the volatilization of silver.

This theory seems plausible. Let us examine it. Let us suppose two furnaces in operation, receiving equal quantities of the same kind of ore. They are alike in every respect, rotating with equal speed, fired in the same manner, and with the same amount of draft. They will yield like results.

We will now suppose that both of the cylinders are rotating at the highest feasible rate of speed, thus affording the greatest exposure of the ore to the roasting influences. It is desired to modify their action so that the time occupied by the passage of the ore through them may be doubled.

In one, which we will call the Howell, this result is attained by reducing the speed of rotation. (Owing to the effect of tangential force the reduction of speed will be somewhat less than half.) In the other, or Thompson, by lessening the angle of inclination. The supply of ore to each is supposed to remain unchanged for the present. An analysis of the effects of these alterations gives the following results.

The quantity of ore passed through each furnace in a given time remains unaltered.

The quantity of fuel and air consumed by each is unchanged.

The quantity of ore exposed to the heat at any given moment in each is doubled; therefore, the time during which the ore is exposed to heat in each is doubled.

The length of time during which the ore is dropping through the air is, in the Howell, unchanged; in the Thompson, doubled.

The power consumed in lifting the ore is - in the Howell unchanged; in the Thompson, doubled.

In the Howell the same quantity of work is performed, on the same quantity of ore, with the same expenditure of power, as before the change. In the Thompson a double quantity of work is performed on the same quantity of ore, by a double expenditure of power. If the ore is as well roasted in the one furnace as in the other, the advantage is, so far, with the Howell. Is it probable that the same result can be produced by the Howell as by the Thompson, with half the work, and therefore half the power? As the result sought is a chemical one, the answer to this question depends on the extent to which the attainment of that result is accelerated by the movement of the ore.

The time occupied by the oxidation of a single particle of sulphuret, exposed to heated air, other things being equal, is in some ratio to its mass. It matters not whether an adjacent particle consists of sulphuret or quartz; the two particles of sulphuret will be oxidized simultaneously, in the same length of time as would one particle, provided the requisite quantity of heated oxygen be supplied. As regards the heat, the two particles will aid each other by their combustion, rendering necessary a smaller" proportionate quantity of fuel than if one particle were sulphuret, and the other quartz.

In a roasting cylinder there is, at any given moment, a certain quantity of air, capable of oxidizing a certain number of particles of sulphuret. To expose a greater number of particles would be useless. To expend power for that purpose would be to waste the power.

If we suppose the quantity of oxygen passing through each furnace to be in excess of the quantity which can be utilized, the feed may be increased in both. When the quantity of air passing through the furnaces is in equilibrium with the ore, so that the oxygen is utilized to the utmost practicable extent, one of two things must be true. Either the Thompson is wasting power in lifting the ore more often than the Howell, or, the ore is remaining in it a longer time than is necessary. If the former, the rate of rotation should be reduced; if the latter, the inclination of the cylinder should be increased. The latter is the most probable, because, although it is not denied that oxidation proceeds to a certain extent while the ore lies quiescent in the cylinder, it is proved, by all experience, to proceed faster when the ore is showered through the heated air, provided, as before said, that a sufficient quantity of free oxygen is present, which, on our hypothesis, is the case in this instance.

If the speed of the Thompson is now increased, the same quantity of ore will pass through it, but the time line or passage should be , or whether the result would be a mean in which the diminution of volatilization would compensate, or more than compensate for the greater consumption of power, is a question which can only be answered by experiment.

Erratum - Page 153, third line, for "speed," read, inclination.

A similar view as to the chloridation, which, in this class of furnace, is more or less effected by the falling of the ore particles through the heated gases evolved by the roasting ore and the salt, would lead to similar conclusions. But the theory of the chloridation is opposed to any greater movement of the ore in this stage than is necessary to ensure equal and thorough heating. Hence it would perhaps be advantageous to have a smaller number of lifters in the circumference of the cylinder near its lower or chloridizing end, than in the central and upper zones. But it is known that, in this class of furnace, a considerable percentage of the chloridation takes place after the ore has left the cylinder, and while it is accumulating in the pit into which the being equal, is in some ratio to its mass. It matters not whether an adjacent particle consists of sulphuret or quartz; the two particles of sulphuret will be oxidized simultaneously, in the same length of time as would one particle provided the requisite quantity of number of particles would be useless. 10 expend power for that purpose would be to waste the power. If we suppose the quantity of oxygen passing through each furnace to be in excess of the quantity which can be utilized, the feed may be increased in both. When the quantity of air passing through the furnaces is in equilibrium with the ore, so that the oxygen is utilized to the utmost practicable extent, one of two things must be true. Either the Thompson is wasting power in lifting the ore more often than the Howell, or, the ore is remaining in it a longer time than is necessary. If the former, the rate of rotation should be reduced; if the latter, the inclination of the cylinder should be increased. The latter is the most probable, because, although it is not denied that oxidation proceeds to a certain extent while the ore lies quiescent in the cylinder, it is proved, by all experience, to proceed faster when the ore is showered through the heated air, provided, as before said, that a sufficient quantity of free oxygen is present, which, on our hypothesis, is the case in this instance.

If the speed of the Thompson is now increased, the same quantity of ore will pass through it, but the time occupied in the passage will be lessened, and the load in the furnace will be reduced. Less power will now be consumed in causing the cylinder to rotate, and the loss of silver by volatilization will be diminished.

Whether the power required would be reduced to an equality with that consumed by the Howell, which would mean that the load in the furnace should be reduced to half that in the Howell, or, in other words the time of passage should be half, or whether the result would be a mean in which the diminution of volatilization would compensate, or more than compensate for the greater consumption of power, is a question which can only be answered by experiment.

A similar view as to the chloridation, which, in this class of furnace, is more or less effected by the falling of the ore particles through the heated gases evolved by the roasting ore and the salt, would lead to similar conclusions. But the theory of the chloridation is opposed to any greater movement of the ore in this stage than is necessary to ensure equal and thorough heating. Hence it would perhaps be advantageous to have a smaller number of lifters in the circumference of the cylinder near its lower or chloridizing end, than in the central and upper zones. But it is known that, in this class of furnace, a considerable percentage of the chloridation takes place after the ore has left the cylinder, and while it is accumulating in the pit into which the cylinder empties itself. Exactly how far this principle may be carried in practice has not been proved.

The objection that rapid rotation of the cylinder causes the separation of a larger quantity of dust from the roasting ore, is too puerile to be worth refuting. The only result is that the auxiliary fire has more to do, and the cylinder less, a clear saving of motive power, or an increase in the capacity of the furnace. One point in this connection is worthy of notice; the dusting almost ceases as soon as the ore begins to "sponge." This is in favor of feeding the dust into the lower end of the cylinder.

Another valuable feature in the Thompson furnace, and which has been heretofore surprisingly overlooked by inventors, is that the heat given off by the roasted ore, instead of being wasted, is utilized in heating the air with which the furnace is supplied; at the same time the ore is cooled sufficiently to allow of its being moistened without loss or inconvenience from the formation of a great volume of steam.

Thompson also adapts the length of the cylinder in such proportion to its diameter, and to the extent of grate surface in the fireplace, as that the heat at the higher end shall be, as nearly as practicable, that which is proper for the initial stage of the roasting. It is a self-evident proposition that a large quantity of burning fuel will heat a cylinder, not only of greater diameter, but of greater length, than a smaller quantity. Hence, if the length of the cylinder is not increased in suitable proportion to the increased diameter, one of two things must happen: either the larger cylinder, by being too short, must waste available heat at its upper end, or, the smaller cylinder, being too long, is not sufficiently heated throughout. As the proper initial heat varies for different ores, the most advantageous length of cylinder for a given diameter will also vary. The usual dimensions are 21, 24, 27, and 30 feet in length, to 32, 40, 52, and 60 inches in diameter, respectively.

As to the question of the best manner of regulating the operation of a roasting furnace of this class, the writer's opinion is, that the cylinder should be adjustable both as to speed of rotation and as to inclination, in order that it may be adapted to the various requirements of different ores. In the White, and in Thomp son's modification, this is the case.

The roasting of ores is a problem in which mechanical and chemical factors are inextricably interwoven. While it is well that theory should suggest and guide experiment, yet, if the facts thus developed oppose the theory, the latter, not the former, must give way.