A.--The main cause of them I found to be the largeness of the space left between the valve plates in this class of pumps, and out of which there is nothing to press the air or water which may be lying there. It consequently happens, that if there be the slightest leakage of air into the pump, this air is merely compressed, and not expelled, by the advance of the air pump piston. It expands again to its former bulk on the return of the pump piston, and prevents the water from entering until there is such an accumulation of pressure in the condenser as forces the water into the pump, when the air being expelled by the water, causes a good vacuum to be momentarily formed in the pump when it gorges itself by taking a sudden gulp of water. So soon, however, as the pressure falls in the condenser and some more air leaks into the pump, the former imperfect action recurs and is again redressed in the same violent manner.
465. Q.--Is this irregular action of the pump the cause of the imperfect vacuum?
A.--It is one cause. Sometimes one end of the pump will alone draw and the other end will be inoperative, although it is equally open to the condenser, and this will chiefly take place at the stuffing box end, where a leakage of air is more likely to occur. I find, however, that even when both ends of the pump are acting equally and there is no leakage of air at all, the vacuum maintained by a double acting horizontal pump with india rubber valves, is not so good as that maintained by a single acting pump of the kind usual in old engines.
466. Q.--Will you specify more precisely what were the results you obtained?
A.--When the vacuum pan was exhausted by the pumps without any boiling being carried on in the pan, but only a little cold water being let into it, and also into the pumps to enable them to act in their best manner, it was found that whereas with the old pump a vacuum of 114 on the sugar boiler's gauge could be readily obtained, equal to about 29-1/2 inches of mercury, the lowest that could possibly be got with the new horizontal pump was 122 degrees of the sugar boiler's gauge, or 29 inches of mercury, and to get that the engine must not go faster than 10 or 12 strokes per minute. The proper speed of the engine was 75 strokes per minute, but if allowed to go at that speed the vacuum fell to 130 of the sugar maker's gauge, or 28-1/2 inches of mercury. When the steam was let into the worms of the pan so as to boil the water in it, the vacuum was 134 at 75 revolutions of the engine, and went down to 132 at 40 revolutions, but rose again to 135, equal to about 28-1/4 inches of mercury, at 20 revolutions.
467. Q.--To what do you attribute the circumstance of a better vacuum being got at low speeds than at high speeds?
A.--It is difficult to assign the precise reason, but it appears to be a consequence of the largeness of the vacant space between the valve plates. When the piston of the air pump is drawn back, the air contained in this large collection of water will cause it to boil up like soda water; and when the piston of the pump is forced forward, this air, instead of being expelled, will be again driven into the water. There will consequently be a quantity of air in the pump which cannot be got rid of at all, and which will impair the vacuum as a matter of course.
468. Q.--What expedient did you adopt to improve the vacuum in the engine to which you have referred?
A.--I put blocks of wood on the air pump piston, which at the end of its stroke projected between the valve plates and forced the water out. I also introduced a cock of water at each end of the pump between the valve plates, to insure the presence of water at each end of the pump to force the air out. With these ameliorations the pump worked steadily, and the vacuum obtained became as good as in the old pump. I had previously introduced an injection cock into each end of the air pump in steam vessels, from which I had obtained advantageous results; and in all horizontal air pumps I would recommend the piston and valve plates to be so constructed that the whole of the water will be expressed by the piston. I would also recommend an injection cock to be introduced at each end of the pump.