488 THE COLLIERY GUARDIAN March 7, 1913. through the protective devices in almost every test; except in some of the “ overgas ” tests, the flames were never less than 2 in. in length, and often were longer, the maximum length being about 4 in. The flames lasted only a fraction of a second and varied in colour from pale blue to bright red. The safety of a motor protected as in these tests appears, says Mr. Clark, to depend upon the efficient dissipation of heat in the gaseous mixture surrounding the motor instead of upon the removal of the heat from the escaping gas by the protective devices. “ The gauzes obviously fail in their cooling functions since they do not prevent the passage of flames; therefore it is probable that even when no puncture occurs, the escaping gases fail to ignite the surrounding mixture by only a small margin, which might be overstepped by a slight variation in conditions.” The poppet probably assists in cooling the escaping gases by discharging them m a manner most conducive to rapid loss of heat, and it also serves to equalise the pressure on each side of the gauzes which would otherwise be torn from their fastenings by the explosion within the casing. Generally speak- ing, the conclusion was that since explosions occurred with poppets open, with poppets closed, with 8’6 per cent, of gas (taken as the most explosive mixture), with “ overgas,” with “ undergas,” with low pressure (12'5 lb.), and with comparatively high pressure (30'5 lb.), it is not possible exactly to define what constitutes the explosive condition for Type A device as applied to the particular motor tested. Type B devices were not tested, as the information gathered from the test of Type A devices pointed to Fig. 3.—Section of Type “A” Protective Device. I i. JP Fig.4.—Mounting of Type “D” Protective Device their almost certain failure. Type C devices were also regarded as being foredoomed to failure, and only two tests were made. The openings in the baffle plates were so large that their cooling effect was negligible, and flames could easily pass through them. With Type D devices, 236 tests were made under 25 different conditions. In this case, the armature shaft carried a fan so disposed with reference to the relief openings as to cause a considerable circulation of air through the motor casing. Flames 4 to 6 in. in length were discharged through the protective devices in all tests in which the motor was running with the fan in place, but no outside flames were observed in any other tests. The duration of these flames was short, and this probably accounted for their failure invariably to ignite the surrounding gaseous mixture, unless, perhaps, the first gases discharged from the casing were cooled by the plates below the ignition temperature of methane, and drove away the explosive mixture immediately surrounding the relief openings, so that the flames which issued later were discharged into dead gas. The value of this form of protection may be entirely destroyed by separating the plates by a small fraction of an inch. It is observed, however, that if the motor were to be designed to operate without a fan, “after-burning” sufficient to destroy the insulation would take place if the rotation of the armature should produce a circulation of gas through the casing; as fast as the gas was drawn in, it was burned at the inner edges of the plates. With Type E device 268 tests were made under 26 different conditions. In no tests were any flames observed to issue from the protective devices; when coaldust was sifted into the devices, however, it was driven out in a cloud, and the explosion seemed to start from a point within this cloud. This probably indicates that the dust was raised to its ignition temperature before it left the casing of the motor, and upon reaching the outer air burst into flame and ignited the gaseous mixture surrounding it. The pressure developed in the motor casing in the test in which there was a “ punc- ture” was less than 40 per cent, of the average maximum pressures produced in 10 other tests made with gas alone, in which no “ puncture ” took place. The results obtained indicate that when dust is present Type E devices do not afford the protection which they are designed to give. In a general discussion of the results, Mr. Clark gives the following tabulation comparing the more important features of the devices and the results of the tests made upon them:— General Comparison of Results of Tests of Five Types of Protective Devices. Type. Character of protection. S. Unoc- cupied space in motor casing. R. Mini- mum area of relief openings. R S Maximum pressure, motor at re t 8'6 per cent, gas, front ignition, 20 degs. C., approxi- mate. Maximum pressure under any condition. Dis- charge of flames. Protection from mechanical injury. Conditions under which “ puncture” took place. Cubic feet. Square inches. Lb. per sq.inch. A Gauze protected by a poppet valve 1-2 5’22 4'35 20'0 470 Yes Unsatis- factory (a) Valves open, 8'6 per cent, gas, front ignition, motor at rest. (b) One valve plugged, 10 per- cent. gas, front ignition, motor at rest. B Unprotected gauze* — — — — — — — — C Baffle plates 0'68 2'0 2’95 26'0 27’5 Yes Good 8’6 per cent, gas, front ignition, motor at rest. D Plates 51 48’0 9'4 11'50+ 34-6 Some- times Unsatis- factory 8'6 per cent, gas, brush ignition, motor running. E Subdivided gauze and baffle plates 0'83 2'8 3'38 16’5 47'2 No Good (a) 8'6 per cent, gas and dust, front ignition, motor at rest. (bl 8’6 per cent, gas and dust, back ignition, motor at rest. * Not tested. + The armature fan, even at rest, increased the effective resistance of these protective devices to the passage of gas. With the fan removed, but other conditions the same, the maximum pressure was only 3 lb. per square inch. ■A'- Fig. 5.—Front and Back of Type “ E ” Protective Device. A weak point was found in each type of protection. In only four tests out of 191 did the Type A devices prevent the passage of flames. Such a condition does not appear to be safe and, although few actual “ punc- tures” occurred, the margin of safety seems to be narrow at all times. The successful operation of this type of protection appears to depend upon the dissipa- tion of heat outside of the devices instead of, as was intended, within them. The margin of safety was overstepped when the maximum heat was discharged from the devices at a comparatively low pressure. This was accomplished by exploding the gaseous mixture within the motor casing while the protecting poppet valves were in the open position. Such a condition might exist in practice for the purpose of cooling the motor while in operation. The failure of the Type D devices was due more to the extremely severe conditions imposed by the motor design than to imperfections in the design of the devices themselves. It is hard to conceive of a device of reasonable size that could absorb the amount of heat produced by the continuous operation of such a motor in a gaseous atmosphere. The weak point of the Type E devices was their tendency to ignite coaldust deposited in the openings of the baffle plates outside the gauzes. Explosion-proof protective devices should be so designed that flames cannot be driven through them by an explosion of gas within the casing of the motor to which they are' connected. They should also be so designed that coaldust cannot enter them, or, if allowed to enter, cannot be discharged while ignited. Such devices should be very rugged in mechanical design, or they should be completely protected from injury. They should also be so constructed and applied to the motor that they cannot be detached from it without rendering the motor inoperative. The casings of explosion-proof motors should be designed with the fewest possible number of openings. An ideal casing for this purpose would be one that had no openings whatever, except those to which the protective devices were attached. The electrical conductors which enter such casings should be efficiently bushed with hard fibre, or some other material that will make a strong and tight joint. If the amount of unoccupied space within Fig. 6.—Dimensions of Type “ E ” Protective Device. ‘7u Drill |1! if i the casing is made as small as possible, the duty of the protective devices will be reduced to a minimum. The factory tests to discover unprotected openings in the casings of explosion-proof motors should be as rigid and complete as any tests made upon the protective devices themselves. Several incidents that occurred during this investigation emphasised the importance of making such tests for tightness. The starting boxes of explosion- proof motors should be protected with the same care as the motor casings. The most satisfactory form of protective device is one that is capable of absorbing a large amount of heat. In order to do this the device must be constructed of a metal that is a good conductor of heat. A considerable amount of metal should be used, and it should be so disposed as to offer a large amount of heat-absorbing surface to the flames without being itself raised to an unsafe temperature. The total area of protected openings through the wall of the motor casing should be as large as is consistent mechanically. There are two reasons for this arrangement. First, it permits the use of more heat-absorbing material, and, second, it decreases the maximum pressure developed. The lower the pressure the lower the temperature of the flame and the less heat is passed through the cooling devices in a