The first installation of skip filling commercially successful on a large scale was made at the Edgar Thompson Steel Works, under the supervision of its furnace superintendent, D. G. Kerr (now first vice-president, United States Steel Corporation), and chief engineer, E. E. Slick (now vice-president and general manager, Midvale Steel Company). By careful, painstaking and long-continued experimentation with model tops built exactly to scale, and raw materials crushed and sieved to the same scale, the proportions were found which would give a reasonably uniform and satisfactory distribution of the stock on the bell. A furnace was equipped with a filling apparatus of this design, and it was so successful that except for minor details nothing else has ever since been used at that plant, a record which I think is better than that of almost any other plant in the country, since many of them have tried out two or three, and others very many more types of tops.

A typical double-bell furnace top is illustrated by Figs. 35, 36, 37 in which the main bell and hopper, the upper or gas seal bell with the hood to connect the seat of the upper bell with the top of the hopper, and all the accessories of the top, including the skip, are well shown.

The upper bell does not need to be nearly so large as the lower one, since the diameter of the latter is controlled by the size of the furnace top into which it discharges the charge, but the charge will arrange itself on the main bell in very much the same way irrespective of the size of the upper bell. In fact small size is of decided advantage, since the nearer the stock enters through a mathematical point in the center of the furnace, the more uniform will its distribution be. If the upper hopper be large, and the stock be delivered into it irregularly by the skips, a greater or less portion of these irregularities will follow through the upper bell down into the lower. The larger the upper hopper the greater these irregularities may be, and the more completely they follow down onto the main bell, and so into the furnace.

Many plants have adopted what is known as the plain double-bell style of skip filling, but many have had most disastrous results with it, and others after the most careful and conscientious trials to obtain an exact distribution, have reached the conclusion that no form or design of this top is satisfactory because the distance which the materials travel is affected by the size, shape and moisture contents of the different materials and by the velocity of the skip when dumping. What is a perfect distribution on a dry day may be a frightfully bad one during a rain storm if the ore becomes wet, while a good distribution for a coarse coke, free from fine material, may give horrible results with a fine coke much broken up.

Various ingenious schemes have been used to eliminate or control these irregularities, and baffle plates without number have been tried and then discarded for the reason that no matter how good the results might be with one condition, they would infallibly be bad with some change in the conditions, such as was sure in ordinary operation to arise within a short time. Two of these attempts may be mentioned as showing the lines of endeavor.

The Roberts Top

The Roberts top designed by Frank C. Roberts of Philadelphia is shown by Figs. 35, 36, 37, showing a central section in the plane of the skip, a front elevation at right angles therewith, and the plan view showing the receiving hopper divided into two triangular pieces, more properly half pyramids, the hopper being a truncated pyramid with its axis in the center line of the furnace, and cut in two in the central plane between the skips. Each of these half-hoppers is mounted on four wheels resting on a stationary track at right angles to the plane of the skip. The treads of the wheels are wide, and their flanges set back well inside the line of the track so that the half-hoppers cannot only be moved in and out, but can also be slewed in the horizontal plane in order to secure that distribution which, under test, seems to give the best results. In other respects the design is similar to that of any other double-bell type of top.