It has been very widely believed by many furnacemen, and is even yet, that one of the essential characteristics of a revolving top is that if the revolving apparatus ceases to work it should still continue to give a reasonably good stock distribution. This means that dumping in heaps is barred because, as explained in regard to the Brown and Baker tops, if they cease to rotate they give the worst conceivable distribution. This then implies that the stock must be dumped in annular layers as nearly uniform as possible, and that only irregularities in the layers due to the irregular discharge by the skips into the receiving hopper should be eliminated by revolving the latter.

This idea seemed to have occurred almost simultaneously to various engineers, but the credit for having first applied it successfully undoubtedly belongs to Arthur G. McKee of Cleveland.

The earliest form of this top consisted simply of a conical receiving hopper into which the skips dumped, terminating in a hopper of steeper slope below whose bottom was closed by an ordinary gas-seal bell operated through a hollow bell rod which encircled the main bell rod. The hopper was mounted on ball bearings set well back from the line of the hopper, and carefully designed to keep the dust out from above, and give it ample opportunity to fall out below, holes being drilled at intervals in the bottom of the ball race to facilitate the latter purpose.

The gas-seal bell, of course, has to rotate with the hopper and therefore the hollow bell rod is provided with a small ball bearing near its upper end to permit its lower part to rotate with the small bell without interfering with the operating mechanism above.

Probably one of the most important features of the design in operation is the extension of the ball track to a path well outside of the closest point on the hopper, which is carried by radial arms extending in from the upper ball race. This makes an air space between the ball race and the hopper so that unless the latter be heated to a red heat or above, the ball race still remains relatively unaffected and able to perform its functions.

Fig. 44. Kennedy top.

This apparatus is not driven by the rotation of the sheaves through ratchet wheel and gears as in the Brown, but is provided with an independent electric motor operating through reduction gears, and controlled from the operating platform at the bottom of the furnace.

The general appearance of this early form of this top is shown by Fig. 44, which, however, is not a McKee top, but a top designed by Julian Kennedy of Pittsburg, on exactly the same principle as the McKee, but rotated by a wire rope passing around the hopper and down to the ground level, where it is moved by hydraulic cylinders or other suitable mechanism to give the desired rotation to the hopper. In principle the design is in all respects similar to the McKee.

In the McKee top contacts are attached to the hopper at four points and by means of wires carried down to the operator's platform these light indicator lamps when they come into certain positions.

In operating this first design a skip was dumped into the hopper, which was then rotated through ninety degrees. The second skip was then dumped and the hopper again rotated, and so on to the fourth skip load which completed the cycle, the upper bell was then opened, permitting the charge to descend onto the main bell from which it was dumped after the upper bell was closed. The process was then repeated.

Almost innumerable modifications of this system were designed to eliminate certain irregularities which briefly are the same as those I have pointed out in connection with the dumping of barrows into the hopper in hand-filling. The location of the first skip load in the hopper and the room it occupies is very different from that of the second and third, and still more different from that of the fourth, which is likewise different from that of the second and third. Moreover, with the double skip there was a serious irregularity which does not appear from the examination of a side elevation of one of these tops. From this it might appear that both skips dumped in the central plane between them, but as a matter of fact one dumps well on one side, and the other as far on the other side of this central plane, so that two successive loads are a considerable distance apart in the hopper irrespective of its rotation. It will be obvious on a little consideration that when this distance is combined with that arising from the constant angle of rotation, two adjacent skip loads may be the sum of these two distances apart, and the next two only their difference, with consequent great variation in the shape and location of the resulting heaps.

Many furnaces equipped with tops of this type, while they apparently had a good stock distribution, gave extremely poor results, and as furnaces working under very different conditions gave similar results along certain lines, which were not nearly so noticeable with any other type of top, it seemed reasonable to assume that the trouble was due to this type of distribution.

The furnaces filled in this way scoured out their linings in a zone twenty or thirty feet above the tuyeres, not on one side as so frequently happened with non-rotating skip-filled furnaces, but quite regularly all around. The blast pressures were high, furnace control unsatisfactory, and fuel consumption apt to be excessive.

The real reason for these results has, so far as I know, never been worked out, but my experience in one case proved that they did exist and that they were directly chargeable to the type of distribution system. The proof of these facts was so convincing and the result obtained by eliminating these conditions so remarkable that they must necessarily carry conviction to the mind of any unprejudiced furnace-man. I shall revert to this phase of the subject later.

It was in order to overcome these difficulties, whatever their reason may be, that Mr. McKee has in recent years redesigned his top so as to eliminate some of the worst irregularities. The new design is shown diagrammatically by Fig. 45.

The original rotating conical receiving hopper is entirely eliminated and is replaced by a stationary receiving hopper triangular in front elevation through which each skip load in turn passes into the cylindrical rotating hopper below. This has the great advantage of practically centering the charge to begin with, and thereby eliminates some of the worst irregularities of the original design.

Fig. 45. Recent design of McKee top.

The revolving hopper is now being rotated preferably to six points instead of four, and so equalizes any irregularity which may occur in the discharge of the skip load into the hopper, by throwing it to each one of the six successive radial positions in turn.

The design of ball races carried well away from the hopper proper, the provisions for preventing the entrance of dust and for allowing it free exit are maintained substantially as they were in the earlier design. It is no more than fair to say in this connection that mechanically this top in my experience gave a minimum of trouble, and I think that this experience has been quite general. I have not operated any furnaces with the improved design, but my information is that many of the troubles which prevailed with the older design have been diminished with this, and that the results are very much better, as is to be expected from the removal of the irregularities due to the varying distribution of the skip loads around the conical hopper.