Lighting ›
Comparison of luminaires in practice: Can the designer rely on the manufacturer's data sheets?
In lighting design, when working independently of any manufacturer interests, one is often faced with the question of which product to use for a desired application. The technical requirements are, of course, defined in the plans but still the question remains: which manufacturer has the best product with the best price/performance ratio? On the market there are numerous similar products whose differences it is very difficult to recognize.
Not only the price but also luminous intensity distribution, luminous flux, power consumption, colour temperature, colour rendering and life cycle are parameters on the basis of which the designer usually makes his decision. However, do these data alone suffice to form a firm basis for a decision? Do the data provided really correspond to the product? From experience gathered in our own accredited laboratory we know that this is, unfortunately, not always the case.
We have undertaken model investigations which reflect our experience in a »luminaire test«. For this purpose we have evaluated one specific product type from four reputable manufacturers in the German-Austrian-Swiss region.
The product under examination is a direct linear LED luminaire for indoor application which can be used as a continuous line luminaire or as an individual luminaire for the illumination of industrial buildings. The products chosen and the manufacturers remain anonymous since it is not our purpose to present a product or a manufacturer as »good« or »bad«. More important for us is the basic question of to what extent we can rely on manufacturers' data and which evaluation measures make sense.
Luminous intensity distribution
Luminous intensity distribution should always be taken into consideration when choosing a product. Good conclusions can be drawn from this about the lighting effect to be expected. According to the descriptions of the manufacturers all of the luminaires have a distribution with the property »double asymmetrical«. It becomes clear when one takes a look at the luminous intensity distribution curve of the tested products that this term tells us very little about the properties of the light beam.

Luminous intensity distribution curves of the tested products
The luminous intensity distribution curve of the continuous line luminaires is symmetric to the C0-C90 planes. The C0-C180 plane has a very broad beam so that, when used in the illumination of shelving in a warehouse, high vertical illuminances are achieved. Nevertheless, there are very great differences within these planes.
Luminaire luminous flux
Information about the luminous flux of a luminaire is important for the dimensioning of a lighting installation. Here the user must, in general, rely on the information supplied on the data sheet or in photometric files provided by the manufacturer. More efficient luminaires have lower luminous flux, on average, than less efficient luminaires. The designer must, therefore, usually weigh up whether a higher efficiency or a higher luminous flux is to take priority.
The following diagram shows the luminous flux of the products when tested in our laboratory and compares the results with the details on the data sheet. It must be taken into account that the data from the DIAL measurements include a measurement uncertainty of 5%. This is shown in the diagram using an error indicator.

Luminous flux of the tested products
Products 2, 3 and 4 are within the range of tolerance, with a deviation from the measured value of a maximum of 7,4 %. Of course, scatter based on production-dependent tolerances caused must also be taken into consideration. In our laboratory only one product of each type was tested.
Nevertheless, it is striking that despite the permitted tolerances the manufacturer's figures are higher for all 4 products than the luminous flux which we measured. This is particularly striking for product 1. Here the information on the data sheet deviates by more than 45% from the luminous flux measured. However, the power consumption for this luminaire is 30% lower which leads to the possible conclusion that the LED modules in the product may be different from those indicated in the data sheet.
Luminous efficacy
To evaluate a luminaire, reference is often made to the luminous efficacy of the luminaire. When making comparisons, this is only permissible when evaluating products with the same correlated colour temperature, the same colour rendering index and, at the very least, similar photometric properties. This was the case in our investigation.
We determined that the power consumption indicated by the manufacturer corresponded in the case of all the products tested to the power consumption measured. The maximum deviation was 4% – with the exception of product 1 which has already been mentioned. The power consumption of three of the four products was lower, of one product slightly higher.
In order to determine the luminous efficacy of the luminaire, the luminous flux measured was related to the power consumption measured. The luminous efficacy of the products tested was between 100 and 150 lm/W. These values correspond to the usual range for LED luminaires of this type on the market at present.
However, it is noticeable that, for all the products, the luminous efficacy on the data sheet was slightly higher than the value which we measured. However, this deviation was, as a rule, within the permitted range of tolerance. Nevertheless, product 1 again was the odd one out in that the measured value deviated by approx. -12% from the value on the data sheet.

Luminous efficacy of the tested products
Colour temperature and colour rendering index
For LEDs, in addition to photometric parameters, the topic of light colour has become more and more important. Frequently the correlated colour temperature is used as a simple metric. For all the products tested the manufacturer had indicated a correlated colour temperature of 4 000 K. When measuring the colour temperature, we found no serious deviations. The maximum deviation was 5% which is absolutely uncritical. However, since it is the »correlated colour temperature« we are dealing with for LEDs, the visual appearance can be very different even if the values are identical.

Colour temperature of the tested products
Even if the colour rendering quality does not play a great role for most industrial applications, we have noticed that the general colour rendering index Ra > 80 given by the manufacturer was achieved by all the products tested.
What we were not able to learn from the colour temperature and colour rendering was what effect the impression of colour had on a surface and how homogenously the colour temperature appeared within the luminous intensity distribution. Even if there are now methods to determine the homogeneity of light colour of a luminaire, these data are very abstract and the value of the information for anyone with experience in photometry is very limited. A sample setup and a trained eye can provide more information than a photometric process. For this purpose we arranged the luminaires at a height of 4.5 m, parallel to and at a distance of 1 m from a matt white wall and evaluated the result visually.

Comparison of the lighting effect of the tested products in the White Laboratory at DIAL
Colour streaks can be clearly identified for products 2 and 4. Of course, there are many applications where this mixture of colour plays only a subordinate role, if at all. Nevertheless, the planner must be aware of the problem, for when illuminating a white wall this could lead to unnecessary and avoidable dissatisfaction on the part of the client.
When testing samples of a luminaire, yet more information can be collected, for example about the uniformity of brightness which it would be almost impossible to describe without photometric simulation. This is also part of the complete evaluation of luminaires. It may be that products which do not stand out particularly with regard to luminous flux and luminous efficacy, provide very homogenous light.
Conclusion
In order to determine how suitable a luminaire is for the application in question, all the relevant parameters relating to this application must be known. »Simple« metrics such as luminous efficacy and colour rendering can be of assistance when searching for a cross-section of luminaires from the countless products on the market which might be suitable for one's own requirements. However, to select products really successfully, further criteria are necessary. The photometric values of the products which we have examined are generally lower than the values indicated by manufacturers in their data sheets even after taking measurement uncertainty and possible production fluctuations into account.
We recommend sample testing of all products, since there are other aspects beyond the metrics already used to describe a luminaire, which cannot really be expressed in figures. It is well-nigh impossible to identify photometrically marks or colour streaks in the light produced by a luminaire, but these can be identified easily by the human eye. The finish of a luminaire or its benefits for installation and maintenance cannot be recorded in figures but play a significant role for large lighting installations.
Data sheets tell us a lot but not everything about a product. Particularly in the age of LEDs, a sample testing of products is important and to be highly recommended. And if a planner wishes to err on the side of caution, especially for large contracts, and does not wish to rely only on the manufacturer's information, he should have the product tested by an independent lighting laboratory.