Three separate sets of samples of the LED make & model being tested must be connected to a suitable power source.
Each set should be tested at a specific temperature, 55ºC, 85ºC and one other temperature that the manufacturer may select. This is done so that data between tests can be easily compared using the 55ºC and 85ºC data sets. The third temperature, set by the manufacturer, is available so that the performance of the LED can be highlighted if it has been designed for a particular application environment such as cold store or oven.
The duration of the LM-80 test must be no less than 6,000 hours.
Measuring the lumen maintenance
At intervals determined by the manufacturer the luminous flux (lumen output) of each LED is measured and recorded. The initial output is deemed to be 100% and all subsequent measures are recorded as a % of the initial output. When lumen output is expressed as a % of an initial value it is termed “lumen maintenance”. For the first few hours of the test it is usual for the lumen output of an LED to increase to perhaps 103-105% of the initial value. Thereafter, the lumen output of each LED will gradually decline. This is lumen depreciation. It is almost always most rapid in the sample that is being tested at the highest temperature.
Lumen maintenance is averaged across each sample and plotted on a graph, as below.
This is a lumen maintenance graph from an LM-80 report.
Lumen maintenance is generally expressed as an L value, so lumen maintenance of 70% or 80% would be abbreviated to L70, or L80. Using the above graph as an example we could say that at 105ºC these LEDs reach L94 after 9,000 hours of operation.
Measuring the chromaticity shift: At intervals determined by the manufacturer the colour of the light emitted by each LED is also recorded and plotted on the CIE 1976 colour space diagram. This enables the colour of each LED, each time it is measured, to be expressed as a numerical value, rather as a grid reference on a map gives a numerical expression of a position.
Then, the magnitude of the changes in position of each LED can be expressed as a number, and all the numbers averaged for the sample. This gives an overall view of how stable (or not!) is the colour of the light being emitted. These data are then plotted on a graph, as below:
Finally, please note that there is no pass/fail level for an LM-80 test. Strictly speaking, LM-80 is not a test at all. It is an approved method of evaluating the quality of LEDs with regard to their lumen depreciation and colour shift at 3 different temperatures over a minimum of 6,000 hours.
LM-80 describes how the evaluation should be conducted and how the resulting data should be presented.
