Determination of the Sensitiveness of Automatic Sprinklers.*
Automatic sprinklers for the extinguishment of fires are intreduced to such an extent now that a scientific analysis of their sensitiveness is a necessity in order to form a correct judgment of their true action. Let it be premised that all modern automatic sprinklers use the same low-fusing solder (155° Fahr.) for holding their outlets closed. Yet the time and temperature at which the solder fuses, which secures the valve in place, varies considerably with different makes of sprinklers. This difference in time and temperature is due to the quantity of solder to be melted, the amount of metal with which the solder is in actual contact, the proximity of the solder joint to large masses of metal, and the strain upon the solder.
The purpose of this paper is to correct the erroneous opinion now held as to the actual temperatures at which automatic sprinklers open.
The method generally pursued to express the sensitiveness of sprinklers is to place the sprinkler to be tested in an oven, heated by a common gas stove, ami note the temperature at which it opens. Such tests, so broadly stated, give rise to very different published statements of their sensitiveness. There arc many details connected with a scientific lest which should be stated with precision in order to make tests made by different parties comparable. Ovens are made of wood, lined with tin, iron and glass ; anti the sizes of ovens vary from i cubic foot to 2400 cubic feet ; and the time during which the test continues is never mentioned. The reading of the mercurial thermometer to a fraction of a degree, however, is regarded as all-important, and taken to be a conclusive and correct indication of the sensitiveness of the sprinkler. The writer having observed these published results for many years, and made many experiments himself, has become convinced that the mercurial thermometer is a very sluggish means for correctly indicating such rapid changing temperatures as are going on within an oven heated by burning gases.
Air carries such a small volume of heat that it must be heated to quite an excess of temperature above that indicated by the thermometer, in order to make the thermometer ascend at a rate corresponding to the conditions of a fire. That air is such a delicate carrier of heat may be best understood when we consider that it contains less than as much heat for the same volume as water. To demonstrate this, a cubic foot of water at 62° Fahr. weighs 62.3 pounds, and a cubic foot of air at the same temperature .0761 pounds, a ratio of 819 to 1 ; ami as the specific heat of air at this temperature is only .2377, the regular volumes of heat contained in the two substances are as 3025 to 1.
A hot-water test would not be delicate enough, for no appreciable difference in the sensitiveness of sprinklers would be discernible by a water-test. Air remains as the only practicable agent, and the one most nearly approximating in its nature to actual fires.
Rut the practical difficulty is in the mercurial thermometer. As nothing else, however, is feasible, the only thing to do is to determine its degree of tardiness. This I have done in the following manner: I use an oven 8 inches square and 24 inches deep, Fig. no, open at the bottom, and a small gas stove placed two inches below it, to send the hot gases upward. The cover joint is 4 3/4 inches from the top, and the thermometer projects 2 inches into the top. Two small vent-holes inch) in the top can be opened or closed to permit of different rates of heating the oven. The sprinkler is screwed into a cross-pipe leaving the solder joint about two inches from the top, or on a level with the thermometer bulb. The heat from the gas stove opens different sprinklers in from 1 to 3 minutes, and at indicated temperatures ranging from 190° to 300°. This time may be considered the same as usually operates in actual fires. Having observed the rise of temperature in the oven every 10 seconds, its varying rate is obtained and recorded in Table A, and graphically illustrated in Fig. III for a slow oven test; and in Table B, and Fig. 112, fora quicker oven test. Column 2 in each case gives the rate, or number of degrees, the temperature increases every ten seconds.
To determine the corrections, the thermometer was heated over the gas stove to 300° or so, and instantly removed, and its fall of temperature noted every 10 seconds in the room which was at 73° Fahr. Some experience was necessary to accomplish this satisfactorily. If the thermometer were allowed to hang still, the air would be heated above the normal temperature of the room (73° and hence no correct knowledge of the surrounding temperature about the thermometer would be had. It became necessary, therefore, to fan the air quite violently in order to maintain the temperature about the thermometer at 73° At first thought this may seem like cooling the thermometer, but that is not so—no fanning can reduce the air below the temperature of the room. The
*Presented at the XXIst meeting American Society of Mechanical Engineers, Cincinnati, May, I890, by A. K. Nagle, Chicago, Ill., member of the society.
Experiments TO DETERMINE TIII-. ACTUAL TEMPERATURES IN AN OVEN FROM THE APPARENT ONES INDICATED BY MERCURIAL THERMOMETER,
I hc thermometer used was a high-grade thermometer. graduated to I° Fahr. and up to 350°. The oven was 8″ X 8″ X 24” deep, with the upper 4 3/4 ” removable, a gas stove placed 2″ below the bottom of the box, anti the bottom open to the air. The oven was made of wood and lined with tin. Thermometer bulb extended 2″ through top.
circulation of the air about the thermometer in the room may also be analogous to the circulation of the hot gases in the oven produced by the natural ascent due to their great heat. Table C (and Fig. 113) give the readings of the thermometer, and column 2 its rate of decrease. Column 3 is the difference between the temperature of the thermometer and that of the room. Column 4 is the mean difference every 10 seconds, and column 5 is the mean difference divided by the amount of fall every 10 seconds.
The factor obtained in column 5 is an important one. is practically a constant 7.80. If multiplied by the amount rise in every 10 seconds it gives the number of degrees to be added to the thermometer readings to give the actual temperature. These results are given in column 3, and arc graphically represented in Fig. 114.
It is evident that the factor 7.80 must depend upon the character of thermometer used. The one used in the foregoing experiment (Therm. A) was graduated to 350°, on a length of 14 inches, and had quite a large bulb. In order to ascertain the different effect on different thermometers, one (Therm. B) graduated to 700°, on 14 inches, and having a comparatively small bulb, was used in like manner as heretofore, with the result as given in Table D, and graphically shown in Fig. 113. In this case the factor 5.44 was found. Although not quite a constant, it is near enough for all practical purposes to make it such. It is evident, however, that in all accurate tests the corrections for tardiness of thermometer should be made in each case.
Fig. 114 is a graphical representation of the amount to be added to the indicated temperature for every probable rate of increase of rise for two thermometers, in order to obtain the actual temperature.
It is certain that temperature alone is not a correct expression of the actual sensitiveness of automatic sprinklers, even if correction be made for the slowness of the mercurial thermometer in responding to the actual increase of the temperature of the hot gases. The area enclosed between the actual temperature line, the time line and the temperature of the room line is, probably, as nearly correct an expression as is possible to obtain. This area might be termed Thermal Minutes, and in this manner each sprinkler could obtain a factor for thermal minutes, which would convey some intelligent idea of its true sensitiveness.
To illustrate the frequent misuse of these temperature tests let us take the case of four sprinklers of two different patterns. A party subjects* his own sprinkler to an oven test where the rise of temperature is such as to open the sprinkler in one minute thirty-five seconds (see Fig. IIII); however, he takes no notice of this time, and reads the thermometer at the precise instant the sprinkler opens, at 216° ; but the actual temperature is about 305°. He then places the sprinkler of the other pattern in the same oven, but owing to the fact that, even with considerable well intentioned care, the oven and the thermometer have not cooled to the same temperature
as when testing his own, the oven now heats up quicker (see Fig. 112), and while the sprinkler opens in one minute sixteen seconds (again he does not note the time), the temperature Is noted with exactness at 238°, but actually about 374°.
Now this extra apparent 22° attributed to this latter sprinkler is a very important and damaging amount with which he goes before the public in selling sprinklers. On the other hand, another party may make just such a test for himself, giving his sprinkler, however, the advantage of a slower test, and find just the same result, but in this case in favor of his device. Now, if the actual thermal minutes had been computed in each case, no appreciable difference would have been found in these two different sprinklers. One would have given 292 thermal minutes, and the other 290. This illustrative case is neither a fanciful nor an extreme one, but such as happens to well-meaning parties. It shows what injury and injustice may be done to innocent parties by reputed scientific tests of sprinklers by underwriters, inspectors and sprinkler manufacturers, ignorant of thermal laws.