In the preceding paper, the authors reported that it was possible to get the spectral energy distributions of light sources or the spectral transmittances of optical filters by solving the matrices of which the elements were the relevant values of retarding potential characteristic curves of a photoelectric tube under illumination of tested lights or transmitted lights through tested filters. In those cases, the solutions were affected remarkably by the errors which were involved in measured values so that solutions which failed to meet the problems in physics or engineering were sometimes obtained. Then, in order to get good results, the only way was to measure the current of photoelectric tube with high accuracy.
In this paper, first, we describe the principle to obtain the spectral energy distributions of light sources by means of linear programming applicable to retarding potential characteristics of a photoelectric tube. Next, we obtain the spectral transmittance of an optical interference filter as a practical example of the method above mentioned.
In computation of it, two sorts of numerical values of the retarding characteristic of the photoelectric tube used in experiment are adopoted say, one is reduced from the theoretical equation and the other is a real measured value. The final results obtained linear programming agree fairly well with those obtained by the conventional spectrophotometric method.
In general, careful efforts are made in experiments to minimize the errors involved in measured values and the least squares method is often used when unavoidable errors still exist. In some cases, however, it may be difficult to repeat the measurements many times or to measure with high accuracy. In those cases, linear programming may be useful for the data processing.
(1) S. Yamaguchi, F. Kobayashi: J: Iluum. Engng Inst. Japan, vol.52, No.4 (1968) 108

View full abstract

Based on the spectral power distributions of actual fluorescent lamps (about 6, 500 K in color temperature), the amounts of line spectra were pre-assigned as 80.0 to 405 nm, 200.0 to 436 nm, 100.0 to 546 nm and 30.0 to 578 nm, respectively.
The existence of the spectral power distributions of the test illuminants with these amounts of line spectra, which fulfil the following requiremnts further was examined.
(1) With respect to any of the reference illuminatnts, the test illuminants have the same chromaticity coordinates and the general color-rendering index of R_a=100, provided that the spectral power distributions of the reference illuminants are normalized as 100 or near 100 at the wavelength of 560 nm.
(2) In addition to the requirement 1), the minimum value of the sum of squares is expected with respected to the difference in spectral power distributions between the reference and the test illuminants
These conditions are reduced to solving the linear homogeneous equations derived by applying the Lagrange multiplier method. The spectral power distributions J_t (λ) of the test illuminant was obtained with respect to the CIE standard illuminant C as the reference, which also satisfied the above two requirements.
However, the computed results give an imaginary spectral power distribution which corresponds to negative radiations in some of the visible spectra. This might suggest the impossibility to realize the general color rendering index of R_a=100 with a fluorescent lamp having the pre-assigned amounts of line spectra shown above.

View full abstract

There are several methods available to determine the coefficient of utilization which is an important factor for predicting illumination levels of an interior lighting installation. As well known, Harrison-Anderson Method, the Zonal Cavity Method, the British Zonal Method, the CIE Method and so on have been used selectively in various countries.
The authors highly appraise the CIE Method among them, because of its theoretical exactness and practical simplicty. However, it had some points left to be improved, particularly the convenience for mechanical computation. They developed an improved method fit for complete computerization, which is called the CIET Method.
In the course of its applications, they surveyed the effect of the implantation of lighting fittngs in order to prepare practical tables of coefficients of utilization, and also to apply the curveplotter for drawing the isolux curves, taking into account both the direct illuminance and the diffused illuminance due to interreflection.
As the result, two computer programs, i. e. one for preparing practical tables of coefficients of utilization and the other for plotting isolux curves in any given condition, were developed.

View full abstract