LED manufacturers always refer to the performance of their LED’s by reference to Board Lumen performance without taking into consideration any losses. They cannot consider any other figure as they have no knowledge of the luminaire type their product will be used in.
The industry standards for measuring LED chip performance is carried out at a LED junction temperature [Tj] of 25°centigrade. The junction temperature refers to the temperature of the base of the LED:
In real applications under an ambient 25° centigrade, it is almost impossible for an LED to operate anywhere near this junction temperature. Nor is it possible to accurately measure the Tj only the PCB temperature close to the solder pad where the LED is placed.
In an industrial LED Luminaire, at an ambient 25°C, the PCB temperature is likely to be in the region of 60°C depending on the thermal management.
Once an LED is placed on a PCB and installed into a luminaire, the performance of that LED at Tj 25°C becomes irrelevant as it will no longer operate at such a temperature and the luminaire lumens become the only relevant factor.
Losses in efficacy will be caused by three factors:
Losses due to a higher operating temperature of the LED vs the stated Tj of 25°C
Electrical Losses from the LED Driver
Optical losses caused by reflectance, optical lenses and clear or diffused covers
When the LED manufacturers measure the efficacy of their products, they use a DC power supply and do not calculate the electrical losses from any power supply.
When you install an LED driver[s] you must calculate the efficiency of the driver and power factor into the efficiency calculation. The driver efficiency can be found on the driver technical specification but will generally be between 90% and 95%. The power factor will also generally be between 90%-95%.
To overall efficiency is calculated by multiplying the driver efficiency by the power factor.
Therefore, if you have a 90% efficient driver and a 0.95 power factor, your overall driver efficiency will be 85.5%
Example of actual losses taken from a BSS LED Low Bay [based on actual testing carried out in 3rd party accredited laboratories in Australia and the UK]:
Osram Duris E5 LED 5000k stating 189 Lm/w as per manufactures data sheet at 25°C Tj
Placed on to a PCB and driven at 65mA using a DC power supply at 25 °C ambient gave 179 Lm/w
Adding an LED Driver, electrical losses need to be taken into account:
Most LED Drivers from manufacturers such as Osram, Philips and Tridonic etc will have an efficiency of around 90% and a power factor of around 0.96. To calculate total electrical losses, multiply the efficiency by the power factor:
0.9 x 0.96 = 0.864 or 86.4%
Therefore, the 179 Lm/w once driven using an LED driver becomes 155 Lm/w
Once the LED panel [PCB] is incorporated into an LED Luminaire [Excluding any clear of diffused cover] you can expect to lose at least a further 10% losses. [The final luminaire LM79 lumens on the BSS LED Low Bay showed a lumen output of 140 lm/w]
Adding a clear cover or optical lenses will reduce lumen efficacy by a further 7-10% and possibly up to 25% when using anything other than a clear diffuser.
In summary, it is our opinion, based on actual testing in certified laboratories, that an LED luminaire performance will be at least 25% lower than the LED manufacturers stated performance.
The addition of optical lenses, clear glass or light diffusing material will add further reduction.
If we consider the highest practical lumens per watt from an LED manufacturer, take for example Seoul Semiconductors who possibly have one of the highest efficacies available at approx 200 lm/w, this would result in luminaire lumens of 150 lm/w before the use of any lens or cover.
Many Chinese LED high bay suppliers quote efficacy of 150 lm/w using Philips Lumiled 3030 2D LED’s. The data sheet produced by Lumiled gives an LED efficacy of 157 lm/w suggesting only a 7 lumen per watt loss. This is mathematically impossible.
Let us consider a manufacturer claiming 170 lm/w from their complete luminaire. A particular example that has a clear diffuser to protect it to IP65.
Working a calculation backwards from 170 lm/w, based on the data above:
170 lumens per watt
Clear diffuser losses at 10% = 189 lumens per watt
Luminaire Optical Losses at 10% = 209 lumens per watt
Electrical Losses at 14% = 244 lumens per watt
Temperature losses from Tj 25°C at 5% = 256 lumens per watt
We accept that there will be slight variations and differences between different drivers, reflectors and optical properties, however, we are simply trying to show that many do not consider the various losses etc or have any indication on how to calculate such data.
Even assuming some errors in these calculations [let's assume 10% error], to achieve Luminaire Lumens of 170 lm/w you will need at least board lumens of 230 lm/w. Who produces an LED in large commercial volumes with this performance?