Experiments with devices constructed by us were carded out partly on a single-cylinder slow-speed machine, partly on a 60horsepower 4-cylinder Maxim engine, and lastly on an 8-cylinder Curtiss engine.
The principle underlying the action of muffling is simple. At the moment of opening- of the exhaust valve the pressure conditions are such that the gases issue at velocities of approximately 2,000 feet per second. The problem is to reduce this velocity below that of sound (1,100 feet per second) without causing undue back pressure. The bark of the open exhaust is due to the issuing of the gases at velocities higher than that of sound, and the main disturbance is suppressed as soon as the velocities are brought below 1100 feet per second. The means at hand to accomplish this are: (a) Cooling of the gases to reduce volume, (b) gradual expansion, (c) internal friction and eddy currents in the gas, and (d) frictional resistance between gases and containers and baffles. Of these the first mentioned is practically negligible, as the degree of cooling can not be great in the time available.
To bring into play the other three means would require a construction having the following essentials: (a) An entrance chamber several times the cylinder volume, to allow of the unrestricted transfer of the gases from the cylinder to this chamber, for the purpose. of preventing undue back pressure, and (b) one or more expanson chambers so provided with baffles as to break up the gas currents in such a manner as to cause decreasing velocity by means of both expansion and friction.
As far as application to the engine is concerned, three solutions are possible. The first is to use individual mufflers for each cylinder. This scheme at first sight has a good deal in its favor, but we found it very difficult to properly stay so many mufflers as to prevent dangerous vibrations.
This article courtesy of http://RacingSecrets.com , your racing technology resource.
The second scheme is to combine manifold and muffler, i. e., to internally construct the manifold to convert it into a complete muffler. This scheme also looks good at first sight; but in order to provide sufficient volume a manifold so constructed would be of large diameter, since the lateral distance between cylinders is restricted, and it is a question whether it is desirable to place so large a heatradiating surface next to the cylinders and the structural members of the plane. The question of adding weight at that height from the base to the side of the cylinders is also of importance with relation to possible excessive vibration. We tried one scheme of this kind on the Maxim engine, as below noted, but with doubtful success.
The third scheme is to use a regular manifold and to connect this by means of flexible hose to the muffler proper. This allows of any convenient placing of the muffler with. reference to engine and to operator. The length of exhaust hose is of no importance, consistent only with low back pressure, and, given this a considerable length of hose or pipe is a positive help to the muffler. We believe on all counts that this combination is the best solution.
The first experimental construction embodied the idea of making the entrance space a centrifugal whirl chamber in which the gases were to lose part of their kinetic energy by mutual interference before passing out through holes or slots in a central pipe. It was made up out of a 4-inch tee, the side branch of which was capped and the cap then bored and threaded for the if inch exhaust pipe, of a 6-horsepower oil engine. The opening from this pipe was placed eccentrically in the cap, so that the gases entered the 4-inch tee tangentially. One of the straight-run openings of the tee was plugged, while the other was bushed to receive a ii-inch pipe. The pipe could be extended varying distances into the tee by a long thread. Tests were made with this central pipe not perforated in any way, but open at the end. The best results were found when this pipe extended to within 1/2 inch of the opposite plug. Later the end of the central pipe was plugged and the pipe perforated with holes, a second one was then tried perforated with slots. In these cases the gas, after whirling around in the chamber found its way into the central pipe through the holes or slots and so on out. All of these devices were only moderately successful and did not seem to promise much.
It was then thought desirable to improve the entrance conditions to the whirl chamber by gradually broadening the entrance pipe, thus introducing the gas tangentially in a wide band. Figure 8 shows the construction. Three concentric central pipes, a 4-inch, a 8-inch and a 2-inch were used, the latter being open to the air at both ends. The central pipes were slotted, the 4-inch and 3-inch pipes as in figure 3-a, the 2-inch pipe as in figure 3-b. It was found that the direction of the slots in the inner pipe caused the gas to move out of one end of the pipe, creating a distinct suction on the other end. This device showed, up much better than the crude first construction. It was estimated to cut out SO per cent of the noise. Back-pressure readings were net taken, as we were as yet mainly interested in noise reduction.
The success of this device on the particularly vicious bark of the oil engine on which it was used led us next to construct a combined manifold and muffler of this type for the Maxim engine. The main features of this muffler were: An expanding, flattened nozzle from each cylinder to lead the gas tangentially into the shell, a common annular chamber between the largest inserted pipe and the shell into which the exhaust from all cylinders first entered, a series of four pipes 2, 3, 4, and 5 inch, with perforations so arranged that the gas must first pass through the wall of the 5-inch pipe in a slanting direction, and having attained a certain velocity in this direction must turn and pass through the 4-inch in the opposite direction because of the slotted openings. Another turn was necessary to pass through the 8-inch, and finally through the 2-inch to the atmosphere.
We did not succeed in getting a. very good idea. of the action of this manifold muffler, as the Maxim engine, which had been loaned us, was recalled. The trial run made promised fairly well, but the end joints between the pipe were not fight and considerable gas escaped. As built, with a cast-iron jacket, this muffler proved exceedingly heavy.
We have not done anything more with the manifold mufflers, but intend to try out one or two other ideas, which, however, have so far reached only the design stage.
The last work was done on an entirely different type from that above described. The underlying idea in this type is to provide an ample receiving chamber, and then as the gases work toward the outlet to provide gradually increasing resistance, until the pulsations are toned down and the gas issues in streams of fairly constant velocity at speeds below that of sound. This should substitute for the bark a hiss like that of escaping steam of low pressure. It might be pointed out that the well-known Maxim muffler is of this general type. Only in this muffler the circular receiving chamber is followed by an annular space surrounding the former, which space is packed with baffle plates of spiral form. The gases here lose velocity by internal friction and surface friction. No attempt is made to utilize the idea of a gradually increasing resistance to the flow toward the outlet.
The scheme was tried out on. the 8-cylinder Curtiss engine. Two manifolds of light-weight steel were cross connected by another manifold, so that we finally had a sinale discharge and only one experimental muffler had to be built. The construction of this is shown in figure 4.
We adopted for this muffler the prismatic form for two reasons. In the first place, it is easier to construct than the circular form; and secondly, we believe that this form may have. some advantages over the circular form for stowing away on an arplane. The essential features are as follows: Gradually expanding entrance nozzle A, a receiving chamber B, two expanding and retarding chambers C and P.
About half of the two expansion chambers are filled with closely packed wire gauze of decreasing mesh, in the following order going toward the outlets:
33 Inches of 2 meshes per Inch -------- 28 ( Layers.)
23 Inches of 3 meshes per Inch -------- 26
2 Inches of 4 meshes per inch---------- 24
1 1/2 Inches of 5 meshes per Inch. ---- 18
1 1/2 Inches of 8 meshes per Inch ----- 20
1 inch of 12 meshes per Inch -----------25
This muffler was connected to the outlet of the cross manifold by a short piece of flexible metallic hose and, in parallel with a quickclosing gate valve. By this means the exhaust could be instantly changed from muffler to open air and hack again. An electric tachometer, carefully calibrated, was used to note-changes in speed, and the back pressures were observed by means of mercury manometers connected to the side manifolds near their connection to the cross manifold.
Trials with this muffler showed the following:
(a) The application of the side manifold and of the cross manifold alone served to tone down the barks considerably. The backpressure observed with only the side manifolds was negligible and with the cross manifold showed about 0.3-inch Hg. at rated output of 70 horsepower.
(b) The application of the muffler raised the back pressure only about 0.1-inch Hg., which is a very good result. The power loss is negligible.
(c) As far as muffling is concerned, three observers judged that the exhaust noise was cut out to the extent of about 50 per cent.
We believe that the muffling efficiency of this construction can be improved, and we have already started to work out an improved desigm The virtual absence of back pressure is the most encouraging feature. Objection has been made to this design on the score of carbon clogging of the wire gauze. Only a service run of some hours' duration can prove this point.
At present we have not used any quantitative scheme of judging degree of noises, but have depended upon several independent observers. This scheme is not wholly successful on account mainly of the other noises present besides the exhaust. As a matter of fact, to get any idea at all of this matter in. connection with the Curtiss engine, it became necessary to extend the pipe through a window and to place the muffler outside of the building. In the University of Michigan tests, above quoted, besides using independent observers, a telephone was used, the observer in a room some distance away noting the distance between himself and the telephone at which he failed to distinguish the exhaust noise. The receiver was placed near the engine. What such a scheme would show in our case is problematical, but we intend to fry it out the coming year. We wish gratefully to acknowledge in this connection the active help of Profs. VI H. Gage and C. A. Peirce, of the Sibley College faculty, and the assistance and facilities supplied by the Thomas Aeromotor Co., of Ithaca, N. Y., and the Curtiss Aeroplane & Motor Corporation, of Buffalo, N. Y.
Further experiments will be made on this type of muffler construction and the determination of the laws affecting feed back pressure power loss are already being investigated and will be the subject of & future report.
