Many different types of burners have been designed and built in the attempt to obtain a flame of known value and capable of reproduction without variation. The lamp which has up to the present time met with approval and is the best known reliable standard is the Pentane lamp. This has undergone several modifications, and the present pattern, intended to represent ten times the candlepower of the Parliamentary candle described above, is the official standard of the London Metropolitan Gas Referees.

The Pentane lamp consists of a steatite ring, or Argand burner, to which pentane vapor is supplied from a flat tank located well above the burner. This tank holds liquid pentane, over the surface of which air passes. The air, becoming saturated with the heavy pentane vapor, falls through a syphon pipe to the burner. The flow is regulated by a tap on the outlet, or a valve on the air inlet of the pentane saturator. No wick is needed. A tall chimney draws the flame up with a considerable draft and insures a steady light. A second outer chimney, concentric with the first, warms the air that is supplied to the flame. No glass chimney is required, but the top of the flame may be inspected for adjustment through a mica window. The top of the flame facing the photometer is shielded by a screen giving a clear opening of 1.85 inches; notwithstanding this, the height of the flame must be carefully adjusted. The total height of the apparatus is about 2 feet 10 inches.

The first satisfactory and practical method of measuring candlepower was the Bunsen method. This method consists in placing a semi-transparent screen of parchment or paper in the path of a line drawn between the two lights, so that the surface of the paper is perpendicular to this center line, and one side of the paper is illuminated by the standard light, and the other side by the light to be measured. A spot in the center of the screen is treated with paraffin to give it a greater transparency. By observing the spot on the paper, or translucent screen, this spot will appear either lighter or darker than the surrounding surface, in proportion as the light thrown against its surface varies. This screen should be shifted until such a distance is reached as shall cause the spot to blend its shade with the surrounding surface. At this point, the screen may be regarded as in balance. The exact point at which this balance occurs is sometimes difficult to determine, but by moving the screen to one side and then to the other until the spot becomes discernible in both directions, a mean position can be established. Owing to the difficulty of viewing both sides of the screen in this method, several auxiliary devices have been devised, by means of which both sides of the screen can be viewed with the full vision. The proper location of mirrors is the usual scheme. This eliminates possible error due to the uneven visual strength of the eyes.

Inasmuch as the law of distance is a computation common to all methods, an illustration of the simple Bunsen apparatus is shown here, figure 212.

Diagram showing fundamental principle of photometer.

Fig. 212. - Diagram showing fundamental principle of photometer.

Let Q be the candlepower of the test light Then the illumination or candlepower value upon the screen, on the test-light side, will be Q/n2. As it is presumed that the screen is in a photometric balance, the illumination upon the opposite side of the screen from the tested lamp will be equal to Q/n2. Let I be the candlepower of the tested lamp, then its illumination upon the screen is I/M2 N and M are distances in feet. But I/M2=Q/N2; or I = M2Q/N2.

Giving values, if the known, or test, lamp has a foot candle-value of 10; the distance N=1.5 feet, and M=5 feet then I =(52x10)/1.52 = 25X10 = 111.1 2.25

If permanent positions are given to the two lights, and the test light is maintained constant, it will be observed that the candlepower represented for any position of the screen can be marked upon the scale, and thus the candlepower of the tested lamp can be read directly without any computations.

Continued experimentation has resulted in improved photometers. One such device possesses much merit in the way of precision and has been largely adapted for both laboratory and practical work in which portable instruments are required. This method is known as the Lumner-Brodhun photometer, and, owing to its almost universal adoption, will be described in detail here.

The illustration shows the essential parts. Light from lamps in the directions A and B falls on the two opposite sides of an opaque screen S. This screen is made of some carefully prepared white material, such as compressed magnesia or barium sulphate. Two mirrors or totally reflecting prisms are placed, at M1 and M2, and a pair of prisms are placed at P1 and P2. These consist of an ordinary right-angled prism P2 with a flat base, and a prism P1, also right-angled, but having part of its base removed by sand blasting or etching. The two bases are brought under pressure into optical contact. The whole arrangement is contained in a blackened box. Light from the direction A falls on the screen S, and is reflected in the mirror M1. Thence some of it passes straight through a telescope to the observer in the direction O while the rest, falling on the sand-blasted portion of the base of prism P1 which is not in optical contact, is stopped.

Diagram of essential parts of Lumner Brodhun photometer.

Fig 213. - Diagram of essential parts of Lumner-Brodhun photometer.

The light from the direction B, shown for clearness by a double line, is reflected from the other face of the screen S, onto the mirror, or prism, M2. Some of it passes straight through the prism P2 and is lost, while the rest falls on that part of the base of P2 which is not in optical contact and is totally reflected to the telescope.

When looking through the eyepiece of the telescope, an unbalanced condition exist: a field of light is observed in which a center spot is revealed in unequal brightness. By moving the apparatus to the proper position between the lights the field assumes a uniform luminosity. The solving for the candlepower of the light to be measured is then performed in the same way as with the simpler Bunsen apparatus.

Diagram showing details of portable photometer.

Fig. 214. - Diagram showing details of portable photometer.

Photometric work carried on at one of the pyrotechnic manufacturing plants was performed with satisfactory results when a portable commercial instrument was used. This instrument was designed by C. H. Sharp and P. L. Millar. Its adaptability to pyrotechnical work merits notice here.

The test lamp A in this case is the adjustable member, instead of the prisms. B and C are two prisms arranged to carry the illuminations from the two sources of light to the telescope F in the same manner as in the Lumner-Brodhun photometer. In this adaptation of the Lumner-Brodhun device, the center or common screen is not present. Rays from the tested light fall upon the diffusing screen E, whose absorption value is known, then reflected by the mirror D to the prism C, and thence to the eyepiece. In case of stronger light sources than that to which the scale indicates, an absorption screen K can be mounted on a swivel device to be thrown into the path of light. The values then indicated will represent a definite per cent. of the true candlepower, according to the known value of the absorption screen. This known value of the absorption screen is a calibrated one and introduces only a negligible amount of error. A still further reduction is available in the reversible elbow piece G D whose surface C is of a diffusing character. The scale reading is directly in candlepower and only requires further computation in case the additional absorbing means are used. In such a case, in the formula, I is the actual candlepower. R is the candlepower reading on the scale and M and N are the per cent values of the light transmitted, respectively, by the diffusion reflector and the absorption screen, and d the distance, in feet, of the tested light from the instrument.

Exterior view of photometer laboratory.

Fig. 215. - Exterior view of photometer laboratory.

The test lamp employed is an incandescent bulb and uses a current and voltage obtainable by the use of four standard dry cells. The lamp is calibrated to be of the correct intensity for the scale. The electric current must be watched closely with an ammeter, and an adjustable resistance.

Figure 211 of the front of the photometer, shows the large knob, used to move the test lamp into balance, as well as the graduated scale indicating the candlepower of the adjusted position of the lamp. The movement of the lamp is effected by means of a non-stretchable cord, which passes around pulleys, and comes to a drum turned by the large external knob.

In the installation mentioned previously for pyrotechnic testing, the apparatus is used to measure the candlepower of aeroplane flares and other pyrotechnic pieces. An external view of the building used is shown in figure 215.

The fireproof enclosure used for burning the pieces to be tested is shown in figure 216.

Fireproof chamber for testing samples.

Fig. 216. - Fireproof chamber for testing samples.

The photometer is located at a distance of 48 feet from the burning piece.

Interior view of photometer laboratory.

Fig. 217. - Interior view of photometer laboratory.

Fig. 217 shows the burning chamber at this 48-foot distance. At this distance candlepowers ranging from 3,000,000 down to 6.4 can be measured with the aid of absorption screens which can be placed in the path of light from the test lamp or the tested pieces.

With certain precautions, results within reasonable accuracy can be obtained. The amperage flowing through the test lamp must be maintained at the tested value. All reflected light must be screened so as not to fall within the field. Adjacent background must be of a dark color and rough of surface. In the case of pyrotechnic pieces proper draft must be maintained to remove the smoke. The full flame of the tested piece should be exposed to the field.

Periodic recalibration of the ammeter and test lamp must be made. When a convenient laboratory is not available a test ammeter and lamp should be provided for this purpose.