In the early days of the closed-top furnace the bell was probably always controlled by a hand-operated "crab," consisting of a crank and connecting rod driven by gearing. A large gear having a crank pin in its face was set immediately under the outer end of the bell beam and a connecting rod ran from a pin in the bell beam to the crank pin in the gear. When the pin was in its lower position the bell was shut, and the travel of the pin was such that when it was in the opposite position the bell was opened wide. A few of these crabs have remained in operation down to quite recent years, but they were superseded in general practice many years ago by a steam cylinder as shown in several of the preceding figures.

There are two methods of using the cylinder. One is to counterweight the bell so that it will stay shut even with the full charge upon it, and have the cylinder lift the counterweight to open it. This has the great advantage that if the steam connections are broken or any other mishap takes place, the bell remains closed and prevents the entrance of air into the top of the furnace, which is necessary in order to prevent gas explosions.

In the other method of operating, the counterweight is dispensed with and the direct downward pull of the cylinder keeps the bell closed. This is somewhat simpler than the other method or at least permits doing away with the counterweight; moreover, by operating in this way the cylinder can be placed at or near the bottom of the furnace and connected with the bell by means of a wire rope, but obviously this cannot be done when the bell is kept shut by a counterweight. However, a counterweight can be applied at the bottom of the rope and immediately over the top of the cylinder if desired.

In some cases a compromise method is used and the bell is counter-weighted about to the extent which will keep it shut when empty but not with the charge upon it. This requires less lift from the cylinder to open than the full counterweight method, and also requires less pull from the cylinder to close than the non-counterweight method. Of course the cylinder is made double acting in this case, which it need not necessarily be in the other cases, although it usually is. By this compromise method, and by equalizing the work on the up and down strokes of the cylinder, the latter can be made considerably smaller than when it is virtually or actually single acting.

The cylinder is an extremely simple method of operating the bell, but is not a particularly satisfactory one. It is obvious that the bell must be held to its seat against the hopper with sufficient pressure to make a gas-tight joint even when loaded with the charge, which probably averages between fifteen and thirty thousand pounds. The amount of actual work involved in opening and closing the bell is very small, but when a cylinder is used for this service, it must be large enough to hold the bell shut or to push it open under the worst conditions which can arise, and the cylinder must, therefore, be much too large for the work it actually has to do most of the time. This is apt to result in violent movement of the bell, unless the cylinder is steadied by a large dash-pot, and in totally unnecessary stresses on the bell beam and its connections. For this reason it is desirable to use some mechanism which has sufficient power to hold the bell tight shut in its closed position without the expenditure of unnecessary energy in moving it. Any mechanism which maintains control over the bell during dumping and closing, so as to prevent excessive velocity is sufficient.

It is very desirable to have a relatively slow movement of the bell at opening and closing so as to prevent slamming and violent inertia shocks on the whole mechanism. For this purpose the bell cylinders are piped as shown in Fig. 43. Inlet takes place through the pipe next the end of the cylinder, but discharge can only take place freely through the one further in from the end, since the outer one contains a check which prevents return flow at any but a very slow rate, so that when the piston overruns the outer port it is cushioned by the steam trapped between itself and the head of the cylinder, and its velocity is thus checked at a rate slow enough to do no harm.

The cylinders generally require to be operated by steam, as compressed air for such purposes is expensive and is likely to be frozen up in winter. When the bell cylinder is placed at the top of the furnace great care must be used to prevent the pipes leading to it from being frozen, since there is no flow except when the bell is operated, and in severe weather during a relatively short suspension of filling an interval sufficient to permit freezing may occur between piston movements.

It will be seen that the crank and connecting rod has an absolutely smooth acceleration and retardation and that in the closed position it acts like an inverted toggle and holds the bell tight shut without expending any energy on it, so that it furnishes the ideal movement for this purpose. To make use of this ideal motion and to dispense with the necessity of steam or compressed air for operating bell cylinders an apparatus has been brought out in recent years which returns to the crank and connecting rod principle; but instead of being hand-operated, as the earlier ones were, it is driven by an electrical motor controlled by automatic switches. The design brought out by the Otis Elevator Company is shown in Fig. 55. Fig. 56 on a smaller scale shows more plainly the arrangement of the crank in relation to the gear. The motor drives worm gears enclosed in the housing shown, and these in turn drive an internal pinion which meshes with the teeth of the annular gear. Automatic switches practically identical with those illustrated in connection with the furnace hoists control the movement of the apparatus. When the starting switch is closed the motor starts and runs until it has driven the main gear and the crank through a complete revolution, when it is automatically slowed down and stopped. This obviously opens the bell and closes it again. The only objection to this apparatus is its very considerable cost, and the necessity of having such an amount of machinery at the top of the furnace.

It may be said that while the cylinder has its disadvantages for operating bells, with certain modifications it could be made to do very much better than the direct-acting cylinder, and in many ways almost equal to the expensive electrically operated apparatus without costing nearly so much.

The operation of gas-seal bells is practically the same as that of large ones, although it requires very much smaller apparatus, since the loads to be carried are much smaller and the small bell itself also weighs many hundred pounds less than the large one. In some cases now the upper bell is operated by an electrical bell hoist similar to but smaller than the one illustrated, for of course it is desirable to have both operated by the same type of mechanism, to have either both steam or both electric.

The upper or gas-seal bells have generally been operated by cylinders working through levers the same as those used on the main bell, but several years ago Julian Kennedy of Pittsburg brought out a design in which the cylinder for the small bell was placed immediately over it in the center of the furnace in an inverted position, its piston rod becoming the rod of the small bell. This, of course, dispenses entirely with the lever for transmitting the motion, etc., but is open to the objection that steam must be on the cylinder practically all the time in order to keep the bell shut, or else a counterweight and lever must be used.

Fig. 55. Electric double screw internal geared bell hoist, direct current.

This design is followed with the rifle-bar cylinder of the Baker top as already noted, but in that case the normal position of the small bell is open, so the cylinder is only under steam when the bell is being raised to close and rotate it.