Bentley T. Barnes

Gas Temperatures and Elastic Losses in Low Pressure Mercury‐Argon Discharges

The increase ΔT in average gas temperature for ac and dc has been determined from measurements of the increase in pressure. A small McLeod gauge was used. Wall temperature was regulated. Corrections for end effects were made by comparing results for long and short tubes of 3.6 cm diameter. Assuming the radial variation of heat input to be parabolic, the temperature distribution was calculated using published values of heat conductivity K. From the results the heat input Pm per centimeter length was calculated. This value was compared with the elastic loss Pc computed from electron temperature Te and number of electrons Ne per centimeter column, taking into account the variation of mean free path with electron velocity. Good agreement was obtained for mercury pressures ranging from 1 to 25 microns, with the arc current 0.2–0.6 ampere and the argon filling pressure 3.5 mm (Hg).

Optical Constants of Incandescent Refractory Metals

Filaments of semicircular cross section (diam 0.25–0.32 mm) were mounted in bulbs with plane windows. Chopped plane-polarized radiant flux was reflected from the flat surface, then passed through an analyzer whose plane of polarization made a 45° angle with the plane of reflection at the filament. The orientation of the polarizer was varied in 45° steps, and wavelengths from 0.47 to 2.0 μ were used. The metals studied were W, Mo, Ta, Ir, Re, Nb, and Pt. Filament temperatures ranged from 300 to 2400–2500°K. (Maximum for Pt was 1900°.) n-vs-λ curves for different temperatures were close together in the visible, but spread apart with increasing wavelength in the infrared. Curves showing k vs λ and e vs λ usually crossed one another near a common point. Deviations from previously published data ranged up to 15 or 20%. Systematic errors in some of the infrared data, due to difficulty in centering the reflecting surface in the beam from the source lamp, are quite obvious. Rough agreement with a simplified form of the Drude theory is approached with increasing wavelength in certain cases.

Intensities of λ1850 and λ2537 in Low‐Pressure Mercury Vapor Lamps with Rare Gas Present

The radiant flux in the Hg resonance lines just inside the wall of a lamp with an inside diameter of 35–36 mm is given for currents 0.4−2.0 amp and various bulb temperatures and rare gas fillings. The intensity of the 1850 line ranges from 12 to 34% of that of the 2537 one, with the highest ratios occurring when both arc current and bulb temperature are at the upper end of the ranges used. Since this result is contrary to what one would expect with single‐stage excitation of the 1850 line, one must assume that this line is excited mainly by 63P–61P1 transitions when arc current and mercury vapor pressure are both relatively high. A rough estimate of the frequency of 63P2–61P1 transitions, compared to those from the ground level to 61P1 confirms this conclusion.

Disturbance of a Low-Pressure Discharge by a Probe

This paper correlates the appearance of the discharge around a probe with data obtained with a fixed and a movable wall probe. A relatively small probe at the axis of a discharge has much less effect than the insulating sleeve around the probe lead. Since ion depletion is the controlling factor, the disturbing effect increases a little as the probe is made more negative. The presence of the insulating sleeve enhances the electric field on the anode side of it, increasing ion production, and displaces the region of maximum ion production toward the part of the tube beyond the end of the probe. Corresponding increases in ion flow to the wall beyond the end of the probe and to either side of it are observed. In noble gas discharges (tube 34 mm i.d., filling pressure 1.0–7.0 mm Hg, are current 1.0–1.4 A) increases in wall current beyond the end of a probe at the axis of the discharge, with an insulator 0.15 mm in diameter around its lead, ranged from 12 to 29%.

Band Systems for Appraisal of Color Rendition

Chromaticity shifts on the RUCS scale, for 20% increase in intensity in a single band 0.01 μ wide, were computed for deluxe cool-white and warm-white fluorescent lamps and ten hypothetical pigments with color points almost uniformly spaced about that of the illuminant. At each wavelength, the color shift of the most sensitive pigment was taken as a measure of the importance of the spectral flux per 100 A. Seven bands of roughly equal importance in rating cool-white lamps have division points at 0.44, 0.49, 0.53, 0.56, 0.60, and 0.62 μ. For warm-white lamps, the second division point listed above should be omitted. The outer limits of the end bands should be 0.42 and 0.67 μ for ratings based on unweighted energy flux. For comparisons of dissimilar spectral curves, one should weight energy flux within each band in proportion to the maximum shift on the RUCS scale, per unit increment of spectral flux. This weighting factor has maxima at approximately 0.445, 0.53, and 0.615 μ and minima at 0.49 and 0.57–0.58 μ.Importance in rating deluxe cool-white lamps is only 5 and 8, respectively (average=100) for Bouma’s end bands, 51–195 for the other six, and 73–136 for Kruithof’s 7-band system.

Techniques for Measuring the Dynamic Characteristics of a Low Pressure Discharge

The dynamic characteristics of a discharge can be determined by making square‐wave changes of arc current. This is done by putting the discharge tube in the plate circuit of a group of pentodes connected in parallel and supplied with a square‐wave signal superimposed on a d.c. bias.Probes operated at constant current are very useful in recording the characteristics of a modulated discharge. For measuring electron temperature (Te), two identical probes are centered about points having the same space potential. If the probes are operated at different fixed currents on the straight line part of the probe characteristic, the voltage between them is proportional to Te. Longitudinal voltage gradient or radial potential differences may be recorded by use of identical probes operated at the same current.A probe operated at a suitable current will follow roughly the variations in space potential of the adjacent region of the discharge. Another probe made sufficiently negative with respect to such a probe will coll...

Determination of Electron Densities in dc Discharges with the Aid of Probes

Oscillation‐free arcs were operated at 1.4 or 1.0 A in a tube with a 34 mm i.d. and a filling pressure of 7 Torr for neon, and 1–2 Torr for Ar, Kr, and Xe. Probes, including one moved radially, were used to determine positive column characteristics, particularly the axial electron density n0. In each computation, corrections were made for the radial variation of gas temperature. The electron density distribution, n/n0 vs r, was not determined with the desired accuracy. Disturbance of the discharge made the probe current rise too rapidly as the probe approached the axis. After due allowance for errors, n/n0 in the neon discharge seemed to vary roughly as the Bessel function J0 (2.33 r/R), where R=radius of positive column. With the heavier gases there was some evidence of a slight constriction of the discharge, presumably due to dissociative recombination of molecular ions formed in the cooler regions near the tube wall. Plausible curves for n/n0 vs r can be drawn to make n0 values computed from the ambipo...

Electron Energy Distributions in Various Discharges

Energy distributions have been derived from the slopes of semilogarithmic plots of the electron current i versus the probe voltage V, and of di/dV. Results are given for dc discharges in inert gases with and without mercury vapor present. Tubes of 34 mm i.d. with filling pressures of 1.0 to 7 Torr and arc currents from 0.2 to 1.5 A were used. At the higher currents, a Maxwellian distribution (MD) was always found for electron energies insufficient for inelastic collision and was sometimes present at higher energies. Deviations from a MD increase with increasing mean energy of the electrons and with increasing distance from the axis of the discharge. The latter effect was enhanced in a tube with neon plus 0.25% Kr in which radial cataphoresis kept the krypton near the tube wall.

Power Losses from Heated Objects in Still Air

The various factors affecting conduction and convection are discussed. Ordinarily, the power carried away from a heated body by these processes cannot be calculated accurately. Two ways of measuring it are described: (1) by boiling water in a container and subtracting from the power input the losses due to evaporation and radiation; and (2) by silvering the bulb of a 500-watt tungsten-filament lamp on the outside, and then making the power input such that the average bulb temperature is the same as that of the unsilvered lamp at normal input. A similar test, but with the bulb silvered on the inside, combined with the results of test (2), gave the power radiated by the bulb. Thus, we find 8 percent of the input to a regular 500-watt lamp carried off by conduction and convection, 11 percent by bulb radiation, the rest by direct filament radiation.Test (1) showed a conduction and convection loss of about 0.05 watt/cm2 at 100°C. This exceeded the radiation loss for every surface tested except a radiator enamel (total emissivity 0.77).