Retrofitting LED Luminaires

Retrofitting LEDs has become widespread over the years, but how clear are you on what design checks are required? Take a look at Guidance Note 6 from the Institution of Lighting Professionals, co-produced with our technical director David Lodge, for a better understanding.

Click here to access > 

Want to learn more about Maintenance Factors?

Want to learn more about Maintenance Factors and how to use them? This free download from the Institution of Lighting Professionals, co-produced with our technical director David Lodge, helps explain how to determine maintenance factors and their impact on the performance and overall efficiency of LED luminaires.

Click here to access >

Safety of LED Luminaires - Blue Light Exposure

Comparison of Blue Light content in LED Luminaires Versus Moonlight and Sunlight

Residents and general public are noticing the change in lighting technology as we move from orange sodium based lighting to white LED lights. The public have raised questions about the relative safety of this technology particularly due to uncontrolled information on the web which is not scientifically rigorous in its approach. This summary provides some clarity based on the scientific data for the safety of LED Luminaires.


Figure 1 below shows the spectral distribution for sunlight. White light we see is made up of a mixture of light wavelength across the rainbow. It also includes spectra outside the visible spectrum including both infra-red (heat) and ultraviolet light which is responsible for tanning skin and when exposed for too long, to an increased risk of cancer.

The mixing of the red, green and blue light creates the impression of white light in the eye.

It is also worth noting that the colour temperature of sunlight (indicating how blue it is) is close to 6500 Kelvin. Note that the higher the colour temperature the bluer the light is and the lower the colour temperature the redder the light is.

Figure 1 Sunlight – Spectral Distribution


Figure 2 shows the spectral distribution for moonlight. This is the baseline distribution for human vision at night. This light has been reflected off the moon back to the earth and so there is some absorption of wavelengths in the ultraviolet spectrum, below 380nm, resulting in significantly less of the light that can lead to damage.

It is also clear that the spectrum needs to consist of similar amounts of red, green and blue light to make the white moonlight.

The colour temperature of Moonlight is 4500K meaning it is less blue than sunlight reflecting the absorption of light in these wavelength by the surface of the moon.

Figure 2 Moonlight – Spectral Distribution

Figure 3 shows the spectral distribution of the LED light involved in the project at Dover. As you can see the peak of the blue light (on the green curve representing our LED) is level with the peak around the red wavelengths of 590nm.

As you can see, the spectral distribution cuts out all of the UV spectrum and the LEDs have less blue light than moonlight because we have reduced the blue light content between 400 and 425nm and between 460 and 525nm. By doing so we are significantly reduced the amount of blue light relative to both moonlight and sunlight, however the mixing of the red, green and blue wavelengths are sufficient to still create a good quality white light.

Figure 3 – LED Light – Spectral Distribution



Although, very few people would argue that exposure to visible spectrum (red through to blue) light is damaging to our health, we know that too much exposure to ultraviolet light does carry a health risk. Let’s look at the cause of the risk from ultraviolet light.
Ultraviolet light is split into three bands, UVA, UVB and UVC.

UVC is furthest from blue light on the electromagnetic spectrum and is the most harmful meaning high energy, light. Fortunately, UVC is completely absorbed by our atmosphere and doesn’t reach our skin to cause us any harm.

UVA light has enough energy to contribute to skin aging but is not high enough energy to cause skin burning that is caused by the higher energy UVB wavelengths.

On a scale of potential harm, UVB is a moderate risk, UVA is a low risk. Blue light from the sun has a very low risk of causing any damage to health and is not recognised as an important factor in causing skin cancer.

However, all UV light has been removed from the LED light spectrum and so we can deduce that the impact on health from the LED light spectrum is consequently very low for equivalent light levels. Which raises our final consideration: how do the light levels from LED lighting compare to sunlight and moonlight levels.


It is difficult to say absolutely what is the quantity of blue light that could affect our health as the impacts are very small meaning very high levels of blue light exposure would be required. Its probably easier to look at the relative levels of light to illustrate the relative impacts of sunlight, moonlight and LED light to give us a feel for the exposure levels versus a common sense position on safe exposure levels.

From the above figures and discussion, it’s clear that all white light we are considering includes a mix of red, green and blue light. We can also see that the proportion of blue light in the LED is much lower than both sunlight and moonlight and that there is none of the more harmful ultraviolet light in the LED light meaning that this is less harmful than sunlight or moonlight at the same light level.

So the question is how much blue light are we exposed to from LED luminaires.

A good measure it to look at how much light (illuminance) we measure as hitting the ground from the three sources. The measure of illuminance is expressed in lux which is a linear scale:


Light Level


Moonlight  < 1 lux  On the ground on a clear night
LED Light 5 lux  On the ground on a clear night at the edge of a lit area
Sunlight  110,000 lux  On the ground on a clear day

It is clear that the light levels that we are exposed to as a result of moonlight are very low, which is intuitively correct. However, the light levels we are exposed to from LED lighting is not significantly higher, being 5 times higher than moonlight at the edges of a lit area. By contrast, the exposure levels from sunlight are 22,000 times higher than from LED luminaires.

Given the ultraviolet light has been eliminated from LED lighting we can deduce that the actual health risks from LED light are even lower this comparison would indicate.


The light level reduces with distance from the LED luminaire by a ratio of the distance squared. For a scheme which is typically lit to 100m from the lighting column or mast, we can show that the light levels for a person stood 150m relative to 100m from the light source are 2.25 times lower. This means the levels at 500m from the lighting mast are half a million times lower or 0.000 002 times the levels in sunlight.


The final evidence is the effects of long term use of other white light sources, such as mercury vapour lamps which were introduced in early 1900s. There have been no health concerns raised with the long term use of these products despite these lamps having a more blue light content and an some ultraviolet light. The evidence suggests that LED luminaires would have relatively less impact on health that these other white light sources due to reduced blue and ultraviolet levels indicating that there is no measurable risk to human health over the long term.


Lighting technology improvements show that LED lighting is both safe and effective having significant safety benefits to those working in the lit area when compared to sodium based light sources.

Similar white light technology (using mercury vapour light) was first introduced in Europe in the early 1900s and no adverse health effects have been recorded from this technology despite the levels of UV being significantly higher than from LED lights we use today.

The refined wavelengths of light from LED luminaires compared with sunlight make LED sources far safer than sunlight. The light levels that we are exposed to at the edge of a lit area are at least 22,000 times less than those we experience from sunlight.

The impact of distance from a lighting installation further reduces the impact. At only 50m distance from the edge of a lit area are less than half those at the edge of the lit area, and at by 500m away from the LED light, the light levels produced are around 500,000 times less than produced by sunlight.

CU Phosco Lighting are a responsible company with high standards of social and environmental responsibility and we only offer products onto the market that are of the highest quality and design. We make every effort to ensure that our products are completely safe both for our customers, there staff and the general public.

David Lodge CEng MICE MIEAust CPEng – Technical Director 
Glaringly Obvious?

Guus Ketelings and David Lodge explain glare, TI% and G-ratings.

It’s late summer, you are sat in the garden, sunglasses on, enjoying a nice cold beverage whilst catching some final late-evening rays as the sun drops towards the horizon - sounds idyllic. But we all know that when driving at this time of day – during sunrise or sunset – glare is often problematic.

Whilst glare in circumstances like this is very noticeable, road users also experience glare during night hours caused by artificial lighting.

Whilst we all know what glare is – what is it really? There are two main types of glare: discomfort and disability glare. Discomfort glare is caused by high luminance in the field of view.  It causes the observer to instinctively look away from the light source. Disability glare impairs the observers’ vision without necessarily causing discomfort. Disability glare is caused by internal reflection of stray light within the eye which is superimposed on the retina. This is called veiling 

luminance.  This in turn reduces the contrast between the observing point and the light source which can affect the ability for the observer to identify hazards or the speed and distance of other vehicles.

The eye needs time to readapt to the surrounding light levels after being exposed to disability glare – especially in low ambience light zones such as rural roads. This can take anywhere from a few seconds to minutes depending on age, health of the observer’s eye. [1]
Note:  the effects of glare reduce as the ambient light levels around the observer increase.  Also, the recovery time relating to the glare reduces with increasing ambient light.


Glare – The basics

On traffic routes, glare is measured in two ways – Threshold Increment (TI%) and Luminous Intensity Classes (G-rating, G0 to G6). Threshold Increment measures the amount of disability glare using a ratio between veiling luminance and background luminance. 

Luminous Intensity Classes on the other hand look at the amount of light within a region above a pre-defined vertical angle. It was originally designed as a measure to categorise luminaire light distributions based on the amount of upwards light produced, with the aim of promoting luminaires that produce less light pollution.  Whilst this parameter does not truly measure glare, assuming the distribution of light is generally directed towards the area to be lit, i.e. the road rather than the verge, it can help indicate the amount of light a driver might experience at different distances from a luminaire.

Luminous Intensity Classes are heavily dependent on the location of the observer relative to the peak beam of the luminaire whereas TI% is heavily dependent on the viewing angle of the observer.

Additionally, Luminous Intensity Classes are omni-directional relative to the luminaire. TI% looks in a fixed direction, and only considers the luminance on the road.

The assessment considers light above the 70° cone with more stringent limitations placed on the regions above 80° and 90°. The reasoning for Luminous Intensity Class being helpful in assessing glare is because of the geometry of the light path from the luminaire to the driver’s eye.  The driver is assumed to be looking at 1° below the horizontal and have a field of view that extends 20° upwards to 19° above the horizontal.  Co-incidentally, light that is emitted above the 71° (90° – 19°) up-angle will be within the -1° to +19° dihedron defining the drivers’ field of view.  However, the background and surroundings will typically be relatively dim resulting in increased contrast and therefore higher disability glare.

A final consideration which isn’t directly addressed through the use of Threshold Increment is the effect of the driver moving along the road. In calculating TI%, the standards assume the driver hops from one calculation grid point to the next. However in reality, the angles of the light to the driver’s eye is constantly changing and where the main beam of the light distribution extend above the 71° up-angle, the driver can experience a flashing effect in the top of the field of view as the luminaire passes out of sight. Selecting a G-Rating with very limited light over the 70 ° up-angle is an effective control and minimises this flashing effect, significantly increasing the comfort for the driver who on long journeys could be exposed the flashing over several hours.  It is likely that a G3 luminaire will create this flashing effect, whilst a G6 luminaire is unlikely to create flashing as can be seen from the green peak beam distributions for G6, G4, G2 and the unclassified polar diagrams below. 


Application of TI% and G-ratings

Recently there has been some discussion around the use of blanket G-ratings on the Highways network and whether they should be used at all. Some say that TI% provides a much better metric for measuring glare specific to the road layout.  Whilst this is true, there are some problems with this.

Both metrics can assist with assessing the potential for glare, as we have established. However, they function slightly differently.

Luminance is only measured for (relatively) straight sections of road that are flat, and so TI% is not easily applied on bends or on the crests of hills. For crests of hills, there is a requirement in BS 5489-1 to use G4 to G6 rated luminaires to limit glare from the luminaires beyond the crest.

While TI% is an important measure of glare, it should be used in conjunction with G-ratings to prevent problems where the calculation of TI% doesn’t fully represent the curves and elevation geometry of the road.


Provided that the chosen luminaire provides you with a suitable distribution, G-ratings (when read in conjunction with the Imax values) provide a good first indicator of potential glare levels. Threshold Increment should then be used to fully and accurately assess glare levels.

Additionally, G-ratings can mitigate problems with glare where Threshold Increment falls short such as on bends or on the crests of hills.

Lastly, increasing G-rating requirements from e.g. G3 to G6 can reduce the amount of apparent flashing from luminaires passing through the top-end of an observers’ field of view.


A point worth considering

Threshold Increment factors in the condition of the eye of the observer – which is assumed to be in perfect condition. For example, the age of the observer’s eye is taken by the standard to be 23 years. However, research suggests that older observers experience higher levels of disability and discomfort glare. This effect levels out at age 50.[2] A large section of the calculations for Threshold Increment were formulated in the 1970s – research was completed before that.

A point worth considering might be that the United Kingdom (and many other western countries) has an aging population. According to the Department for Transport[3], 87% of the population within the 50-59 age band hold a full driving license, with the 40-49 age band (85%) and 60-69 age band (83%) following closely behind. On the contrary, in the 20-29 age band, only 65% hold a full driving license.

Whilst this cannot be taken as a direct extrapolation of how many people from a certain age group use the roads and therefore road lighting, it is clearly demonstrating the gap between the assumptions and real life.

Should we as industry professionals be considering the effect of glare on an aging population and therefore assume that the driver is older than 23 years old?

Guus Ketelings – Lighting Design Technician, David Lodge CEng MICE MIEAust CPEng – Technical Director 


[1] [Van Derlofske J, Chen J, Bullough JD, Akashi Y. 2006. Headlamp glare exposure and recovery (SAE paper 2005- 01-1573). Society of Automotive Engineers 2005 Transactions Journal of Passenger Cars - Mechanical Systems 114(6): 1974-1981.]
[2] [N Davoudian MArch PhD MSLL, P Raynham BSc MSc CEng FILP MCIBSE FSLL and E Barrett MSci PhD. Disability glare: A study in simulated road lighting conditions.]
[3] []