What's New

Lumens per watt claims, can they really be believed?

By May 18, 2016 No Comments

How accurate is the data we are given?

How can you trust published lumens per watt claims?

The LED Lighting industry seems to be judged on performance and in particular, lumens per watt as the most important factor. Lumens per watt are one of the factors you need to consider when making purchasing choices, but many claims are totally exaggerated and impossible to measure without specialist equipment.
The higher the actual lumens per watt, the more efficient the light is and the lower the energy operating costs. However, that’s not the whole story;
There are also many LED brands that will give a very high initial lumen per watt figure but then see 20%-30% lumen depreciation within a very short period of time. Other LED brands will give lower initial lumens but maintain a very low lumen depreciation, sometimes as low as 3% over the same time period.
We also have to consider other items such as CRI [light quality], CCT [colour temperature], thermal management and light distribution etc.
Measuring lumens per watt is achieved by using specialist equipment such as an Integrating Sphere or a Gonio Photometer, these are both expensive pieces of equipment and need specialist training to operate correctly. Many Far East LED manufacturers and assemblers have their own Spheres in their factories and use these to produce documents and certificates to prove their lumens per watt results together with CRI and CCT figures.

As part of an exercise to show how such results can be manipulated, we spent some time in a factory in Ningbo China using a 2000mm integrating sphere to demonstrate how a massive variation on results can be achieved and potential purchasers can be deceived into believing lumens per watt figures.
I would start by saying that, I am not professionally qualified to operate this equipment but I have been fortunate to have worked with professionals within the industry worldwide where I have gained substantial knowledge of the subject matter. I would like to thank, in particular;

  • Richard Hayes and Ian Taylor of 42 Partners for their constant tuition and mentoring
  • David Ford of LED Labs in Sydney Australia for his input and a bottle of beer after a hard day in his laboratory
  • Ken Ball of Light Lab Photometric in Birmingham for proving that a photometer is the best way of achieving accurate rest
  • Barrie Vesty of Auraled in Ningbo China for use of his integrating sphere.

Calibrating the Sphere using a ‘standard lamp’

By using a calibrated test lamp with a known output, the sphere can be checked for accuracy.
In this particular test, the test lamp was certified at 784 lumens at 2.14 amps at 24 V DC [51.36 watts] This provides a simple check that the equipment is recording correct data.

Test Results for the standard lamp showed this to be 14 lumens lower [1.7% difference]

It is also worth noting that the 2000m Sphere, was supplied with a E40 lamp holder and 1500mm LED/Fluorescent tube holder with a baffle to prevent the Light Sensor from direct contact with the light source. The light is reflected around the internal sphere and the light sensor measures the amount of light.

There is also a further adjustment required to take into account the physical luminaire size and how this affects the final reading. On our tests this adjustment was not factored into the results.

Test 1:

Generic 150w LED High Bay

LED Brand: Unknown

Driver: Meanwell HBG 160


There are many who would argue that this product is not suitable for testing in such a sphere, but this whole documents is intended to show how figures can be manipulated by positioning of the luminaire within the sphere to give varying results and a report issued that will serve to prove the lumens per watt claims.
The correct way to mount the luminaire is vertically ensuring that the light sensor within the sphere is protected from direct contact from the light source by the baffle plate.

Test results: 120 lumens per watt

Test 2:

Generic 150w LED High Bay

LED Brand: Unknown

Driver: Meanwell HBG 160

Same product as Test 1 but mounted on a steel table with the light source pointed directly into the light measurement sensor.

Test Results: 166.2 Lumens per watt

Test Summary:

Both tests were recorded and individual test reports were printed.

Therefore, it is proven that by simply manipulating the position of the luminaire in relation to the light sensors, substantial variations in lumens per watt can be achieved and a simple test report/certificate can be produced.
Based on this, it is impossible to rely on data that has been produced anywhere other than a 3rd party accredited laboratory and claims of lumens per watt have to be verified independently by qualified and accredited labs. This isn’t intended to suggest that all non-accredited testing facilities are not capable of producing accurate data, but to allow the industry to realise that it is very simple to manipulate data to give whatever figure is needed.
During other testing, we failed to see any product achieve its published/claimed lumen per watt figure and many showed a 20% plus reduction.


Lumen Depreciation and LM80/TM21 Data

In addition to the initial lumens per watt claims, it is important to know how the LED will perform over time and the expected degradation of the lumen output over the expected life of the luminaires.
At BSSLED we have been very vocal over our warranty and the fact that we cover lumen depreciation as part of the 5-year product warranty. That means that if the lumen output drops below 70% of the original output during the 5-year warranty, we will replace the product under warranty.
We were the first [and possibly still the only manufacturer] to have our LED Low Bay tested to IES LM84 which is a 6000-hour long term lumen test of our complete luminaire where the results showed a reduction in lumens of just 3% over the 6000-hour period.
As a manufacturer, we use Osram Opto Semiconductors as our LED supplier and whilst we appreciate that there are other LED manufacturers with higher initial lumen outputs, we believe in providing a product solution that will perform over its designed life.
What is the benefit of a luminaire producing an initial 150 lumens per watt on day one, when it will lose 25% of its lumen output over the first 6000 hours and then continue to suffer continued degradation during future years until it becomes unfit for purpose?
Many of these LED manufacturers have questionable LM80/TM21 data and a refusal to publish test report data. Just because the data sheet says it is a certain brand of LED, how are you going to verify this data and ensure you are getting what the specification says? Why do 95% of LED luminaire suppliers/manufacturers refuse to disclose the LED brand used in their product?


Without any doubt, the tests carried out over 3 days showed that every product, except the high bay, had exaggerated claims of lumens per watt. It is not our role to publish this data as this would be unethical, but we have raised this matter with the Lighting Industry Association [LIA] in Telford as some of the products were supplied by LIA Members.
The high bay we tested from Auraled LED has a published lumen output of 100 lumens per watt were we achieved a minimum of 110 lumens per watt.
What is clear, is that this is a totally unregulated industry where exaggerated claims are abundant and the authorities do not have the budget or inclination to monitor or police such claims
Ofcom recently told us that they only follow up complaints of interference rather than police EMC compliance.
The LIA face a difficult position balancing membership revenue -vs- compliance but, in my opinion, are following the revenue route.
Having spent years being transparent over our components and providing 3rd party verified products, of course we are going to be vocal and will continue to do so.

Steve Bell
May 2016