During my professional career in lighting, one of my responsibilities was releasing the photometric data for the market for locally manufactured products. Releasing the photometric data, especially with Metal halide lamps, was always a tricky proposition, as the luminous flux changes appreciably when the luminaire is tilted. The lighting design engineers who use this data to present the lighting level to customers (they want to present higher figures to hit competition), always had a different story to tell. They always felt that the data was fictitious and was never achievable in the field. On one occasion, an expert from Europe visited our factory, and during the question/answer session, the aggrieved sales engineer raised a question as to why the published photometric data was never realized in the field and how to correct the state of affairs. He explained by saying that while in other branches of science one plus one is equal to two, in photometry it could be nearly three also, and went on to explain the reason for coming to such a conclusion.
To begin with, in the measurement of length, mass, time, and temperature, the echelon one standards, which are maintained by the various internationally accredited labs, measurement deviations were in a very small band of uncertainty. Success in Space research has proved that humans have mastered these measurements with a very high degree of accuracy. In the case of photometry, candela being a derived standard, when the reference lab results were cross checked for variations, the degree of uncertainty was quite significant as compared to other parameters (Length, mass etc) and we are living with this fact of life.
Also there are many other parameters which can influence the field results from the published values. They are:
1) Variation in supply voltage, the quality of supplied voltage (third and fifth harmonic)
2) Frequency variation.
3) Variation in current setting of ballasts (approved by IEC) and effect due to heating (current value comes down).
4) Variation in lamp lumens during production and the location of support wire within the lamp.
5) Variation in mounting condition (Especially metal halide)
6) Variation in reflective qualities of aluminium mirror from batch to batch.
7) Variation in light absorption of diffusers.
Photometric tests are always done in a dark room at a temperature of 25plus/ minus 2 deg, RH of 65%, and no drafts of air. They always use the same set of reference lamps and ballasts. These reference ballasts have very low temperature rise, and the current characteristics do not vary with time (to a large extent). In the field, if you try to interpolate the published data with actual measurement one would definitely find differences. Even field luxmeters are not properly calibrated. It is always advisable to use correction factor to the published data, based on ones experience, as there are no readymade charts available. It could vary from country to country, place to place, and designers should be aware of these anomalies.
Written by: R. Sitaraman, Director, AWA