How to Mitigate Health & Safety Risks from LED Lighting

Thomas Kaufmann 
Arizona State University 
December 8, 2022 

Abstract

LED light fixtures are the most commonly sold type of light fixture today. Many of these light fixtures flicker, whether visibly or invisibly, and as a result cause adverse health effects, including epileptic seizures, aggravation of autistic behaviors, migraines, headaches, eye strain, and more. The IEEE 1789 standard includes recommendations on how to mitigate potential health and safety risks from LED lighting. This page provides a practical summary of the standard and how to apply it in day-to-day life.

Introduction

Almost half of all light fixtures sold globally in 2019 were LED light fixtures (Placek, 2019). This number is expected to increase to 87% by 2030 (Placek, 2021). Unfortunately, many of these light fixtures have been poorly engineered and as a result exhibit large amounts of flicker. Although for many people this flicker is invisible, it can pose serious health and safety risks, including epileptic seizures, and reduce visual performance.

Many of these light fixtures that can pose health and safety risks are in circulation today. Many facility operators are unaware that their lighting choices may be negatively impacting the health of their building occupants and visitors. At the same time, most individuals who are affected by the adverse health effects of flicker are equally as unaware that invisible flicker from light fixtures or even some video projectors may be the cause of their ailments. In 2008, the Institute of Electrical and Electronics Engineers (IEEE) formed a working group to study these issues and to develop an international standard that provides guidance on how to mitigate these risks. The standard, IEEE 1789, with the title “IEEE Recommended Practices for Modulating Current in High-Brightness LEDs for Mitigating Health Risks to Viewers” was published in 2015 and contains the vast majority of the information presented in this document (IEEE, 2015).

Still, many LED light fixtures available in the market today do not comply with IEEE 1789 recommended practices and the number of environments in which sensitive population groups experience adverse health effects keeps increasing steadily. This document’s purpose is to provide a practical summary of the IEEE 1789 standard and how to practically apply it in day-to-day life.

This page aims to answer the following questions: 

  1. What is flicker?
  2. Why is flicker a problem?
  3. How can we quantify flicker?
  4. How can we avoid adverse health effects from flicker?
  5. What are today’s best practices when it comes to flicker?
  6. How can we measure flicker?
  7. How can we avoid flicker in future lighting purchases?

What is Flicker?

Flicker is defined as “variations of luminance in time”, in other words, changes in brightness or light intensity over time. Flicker can be categorized as follows (Wilkins et al., 2010):
  • Sensation: The eye, brain, and/or neurological system detect the light variation and neurons respond.
  • Visible Flicker: The light variation is sensed and consciously perceived.
  • Invisible Flicker: The light variation is sensed, but not consciously perceived.

In typical light fixtures, flicker tends to be periodic and related to the electrical mains frequency, which is either 50 Hz or 60 Hz. For most people, flicker that occurs at a frequency of 60 Hz or less is directly visible. However, even for flicker frequencies that are invisible to most people, adverse physiological effects can arise. While LEDs themselves do not flicker per se and are not the cause of the problem, flicker can be caused by the way LEDs are powered or dimmed. This phenomenon is entirely avoidable via inexpensive engineering approaches. However, many manufacturers are either unaware of this issue or do not consider it a priority.

Figure 1: Image of a ball bouncing in flickering light (left), demonstrating the stroboscopic effect and image of a vehicle with flickering tail lights (right), demonstrating the phantom array effect (Miller & Lehman, 2015).

Even flicker that is not visible when looking directly and steadily at a flickering light source can become visible via the stroboscopic effect or the phantom array effect (Miller & Lehman, 2015). The stroboscopic effect occurs when the observer’s eye is still, but an object moves through flickering light. The object then appears as multiple discrete copies, as shown in the left image of Figure 1. The phantom array effect occurs when a flickering light source remains still in space, but the observer’s eye moves rapidly across the light source. In this scenario, the light source appears as multiple discrete copies, as shown in the right image of Figure 1. Taking both of these effects into account, it has been shown that humans perceive flicker at frequencies as high as 2 kHz (Davis et al., 2015; Roberts & Wilkins, 2013). Flicker, the stroboscopic effect, and the phantom array effect are considered temporal light artifacts (CIE, 2020).

Why is Flicker a Problem?

Both visible and invisible flicker have been shown to cause adverse health effects in a variety of population groups (IEEE, 2015). Visible flicker at frequencies below approximately 70 Hz is known to cause seizures in individuals with photosensitive epilepsy. However, some photosensitive people may not have received a diagnosis and are unaware that they are at risk. Adverse effects from visible flicker are typically felt immediately.

Invisible flicker at frequencies between 70 Hz and 200 Hz or, when accounting for the stroboscopic effect and the phantom array effect, even up to 2 kHz has been shown to cause fatigue, blurred vision, eyestrain, migraines, severe headaches, nausea, visual disturbances, reduced performance on visual tasks, difficulty reading, increased autistic behaviors, especially in children, apparent slowing or stopping of rotating machinery as a result of the stroboscopic effect, increased distractibility, anxiety, panic attacks, increased heart rate, and dizziness. Invisible flicker can also interfere with imaging devices, such as video and security cameras. Adverse effects from invisible flicker are typically felt after an exposure time of several minutes, but may be felt immediately by sensitive individuals.

How Can We Quantify Flicker?

While there are several different ways to quantify flicker, the two most important metrics are flicker frequency, f, and percent flicker, also known as modulation percentage, Mod%. Flicker frequency describes how often per second a light turns bright and dark. For example, if the flicker frequency is 120 Hz, it means that the light turns from bright to dark and back to bright 120 times per second. The slower the flicker frequency, the more perceptible the flicker becomes. Modulation percentage describes how dark the light becomes compared to the moments when it is at maximum brightness. 0% flicker would mean that the light remains at constant brightness at all times and has no flicker. 100% flicker would mean that the light effectively turns off completely in between the moments of full brightness. The higher the modulation percentage, the more perceptible the flicker becomes.

Figure 2: Example waveforms of three different flicker modulations. The green waveform has a modulation percentage of 10%, which is typical for a standard incandescent light bulb. The yellow waveform has a modulation percentage of 35%, which is typical for poorly designed fluorescent lights. The red waveform has a modulation percentage of 100%, which is not uncommon for poorly designed LED lights.

Figure 2 shows the waveform of three flickering lights. The horizontal axis shows time, the vertical axis shows luminance or brightness. Flicker frequency is defined as how many peaks and valleys of the waveform fit into one second. In this case, all three waveforms have a flicker frequency of 120 Hz. Percent flicker or modulation percentage is defined based on the maximum luminance or brightness, Lmax, and the minimum luminance, Lmin, as follows:

Mod% = 100% × (Lmax – Lmin) / (Lmax + Lmin)

The green waveform shown in Figure 2 has a modulation percentage of 10%, which is typical for a standard incandescent light bulb. The yellow waveform has a modulation percentage of 35%, which is typical for poorly designed fluorescent lights. With poorly designed LED lights, however, even a modulation percentage of 100% is not uncommon, which is shown in red. While this metric does not account for shape and duty cycle of the waveform, it is the most practical metric in the context of this document, particularly to identify light sources with unacceptable amounts of flicker.

How Can We Avoid Adverse Health Effects from Flicker?

IEEE’s working group performed extensive research to identify under what circumstances health risks arise from flicker, how severe the risks are under different conditions, and how to avoid them. Ultimately, IEEE 1789 states three “recommended practices,” each with a distinct goal. The simplest one is “Recommended Practice 3,” whose goal is to prevent photo-epileptic seizures. The recommendation states:

“For any lighting source, under all operating scenarios, flicker Modulation (%) shall satisfy the following goal: Below 90 Hz, Modulation (%) is less than 5%.”

Recommended Practice 1 and 2 are more complicated though. The first question to ask is what the goal of following each of the practices would be.

Recommended Practice 1 should be followed “if it is desired to limit the possible adverse biological effects of flicker.” In the context of this recommendation, the term “low-risk level” is used in IEEE 1789, which is defined as:

“the value of an influential parameter corresponding to a transition between presence and absence of an observable effect in a subpopulation assuming ‘worst-case’ values of other influential parameters.”

In other words, based on available data, even if under certain conditions within these conservative parameters flicker can be observed, it is unlikely to harm anyone. At the same time, the document admits that “the degree of conservatism [...] is not known, although there is consensus among the authors of this document that this contains the low-risk region.”

In contrast, Recommended Practice 2 should be followed “if it is desired to operate within the recommended NOEL of flicker.” NOEL stands for “no observable effect level,” which is defined as:

“an exposure level at which there are no statistically or biologically significant increases in the frequency or severity of any effect between the exposed population and its appropriate control.”

In other words, if Recommended Practice 2 is followed, essentially nobody is harmed in any meaningful way whatsoever. Aside from that, even “annoying” effects from flicker are fully eliminated since flicker at levels within Recommended Practice 2 is essentially not observable under any conditions.

Since there are currently no legal requirements in the United States regarding flicker in lighting, it is for the facility operator to decide whether their desire is to:

  1. Only prevent photo-epileptic seizures, in which case Recommended Practice 3 should be followed,
  2. Reduce adverse health effects from flicker, in which case Recommended Practice 1 should be followed, or
  3. Eliminate adverse health effects and annoyances from flicker altogether, in which case Recommended Practice 2 should be followed.

What Are Today’s Best Practices When It Comes to Flicker?

Recommended Practice 1 of IEEE 1789 states:

“If it is desired to limit the possible adverse biological effects of flicker, then flicker Modulation (%) should satisfy the following goals:

    • Below 90 Hz, Modulation (%) is less than 0.025 × frequency.
    • Between 90 Hz and 1,250 Hz, Modulation (%) is below 0.08 × frequency.
    • Above 1,250 Hz, there is no restriction on Modulation (%).”

Recommended Practice 2 of IEEE 1789 states:

“If it is desired to operate within the recommended NOEL of flicker, then flicker Modulation (%) should be reduced by 2.5 times below the limited biological effect level given in Recommended Practice 1:

    • Below 90 Hz, Modulation (%) is less than 0.01 × frequency.
    • Between 90 Hz and 3,000 Hz, Modulation (%) is below 0.0333 × frequency.
    • Above 3,000 Hz, there is no restriction on Modulation (%).”

While this information may be difficult to comprehend verbally, it is much more elegantly communicated visually.

Figure 3: Diagram showing flicker frequencies and modulation depths within the low-risk region (yellow) and the NOEL (no observable effect limit) region (green) according to IEEE 1789.

Figure 3 shows three different areas. The green area includes flicker frequencies and modulation depths that comply with Recommended Practice 2, which eliminates all adverse health effects from flicker. The yellow area includes flicker frequencies and modulation depths that comply with Recommended Practice 1, which is considered the “low-risk” area. The white area includes flicker frequencies and modulation depths that do not comply with either Recommended Practice and pose a high risk of causing adverse health effects and safety concerns.

In summary, given that flicker frequencies below 90 Hz are unlikely to occur in the first place, the primary formula to consider when attempting to comply with Recommended Practice 1 (RP1) is:

Mod% ≤ f[Hz] × 0.08

The primary formula to consider when attempting to comply with Recommended Practice 2 (RP2) is:

Mod% ≤ f[Hz] × 0.0333

Percent modulation is calculated as discussed above with the formula:

Mod% = 100% × (Lmax – Lmin) / (Lmax + Lmin)

Given that most LED light fixtures flicker at twice the electrical mains frequency, which is either 50 Hz (Europe) or 60 Hz (USA and Canada), the maximum allowable percent modulation in each scenario is as follows:

Electrical Mains Frequency 50 Hz 60 Hz
Typical Flicker Frequency 100 Hz 120 Hz
Maximum Mod% for RP1 (good) 8.0% 9.6%
Maximum Mod% for RP2 (best) 		3.3% 4.0%

How Can We Measure Flicker?

A variety of flicker meters is commercially available today, both in the form of bench-top units as well as portable units. In the context of this document, a handheld device would be strongly preferred to be able to identify flickering light sources in their current location. The United States Department of Energy conducted a study in 2018 to evaluate eight different handheld flicker meters (Leon et al., 2018). The study found that four of those devices produced reliable results. These devices are as follows:

  • Viso Systems LabFlicker
  • UPRtek MK350N Premium
  • GL Optic Spectis 1.0 Touch
  • Gigahertz-Optik BTS256-EF

Unfortunately, these devices are not accessible to many individuals and organizations, as the price of each of these devices exceeds $2,000. At the same time, the functionality of each of these devices far exceeds the requirements of a simple flicker meter.

 

Figure 5: Photograph of Opple Light Master 3 held in a hand (left), screenshot of Opple app risk assessment (middle), and screenshot of Opple app raw data (right). 

An excellent alternative to these options is the Opple Light Master 3. This device connects via Bluetooth to a smartphone and is available for less than $60 here. It provides accurate flicker measurements, a risk assessment based on the IEEE 1789 standard, as well as access to the raw data in the same way an oscilloscope would provide. Figure 5 shows a photograph of the device as well as screenshots of the Opple app on an iPhone.

Another simple way to determine whether a light fixture or video projector creates invisible flicker is to record a “slo-mo” video with a smartphone. (It is important to ensure that the phone is set to the highest frame rate possible, which is 240 fps for recent iPhones. See Settings → Camera → Record Slo-mo → 1080p HD at 240 fps.) Once recorded, the video playback will reveal whether the light source is problematic. If any flicker or strobing effect is visible in the recording, the light source is most likely causing adverse health effects. While this method does not measure flicker in a quantifiable way, it is a fast and easy way to screen for a facility’s worst performing light sources.

How Can We Avoid Flicker in Future Lighting Purchases?

Non-flickering LED light fixtures are generally not more expensive than flickering fixtures. It is simply a matter of carefully selecting the right product. The following strategy should lead to the successful purchase of lighting products that do not cause any adverse health effects due to flicker:

  1. Consult the product specifications of the light fixture or light bulb you plan to purchase. Look for information regarding both flicker frequency and percent modulation. Use those two metrics together with the formula above to calculate whether the light fixture or light bulb complies with Recommended Practice 1, or better, Recommended Practice 2 of IEEE 1789.
  2. If the product specifications of the light fixture or light bulb you plan to purchase do not contain any flicker metrics, consult your lighting vendor and ask whether the product complies with Recommended Practice 1 or Recommended Practice 2 of IEEE 1789. If your lighting vendor does not know the answer, have them consult the manufacturer. If they do not understand the nature of your question, feel free to share this document.
  3. If the first two steps do not lead to success, purchase a flicker meter, such as the Opple Light Master 3. Purchase one of the light fixtures or light bulbs you plan to purchase and test it yourself.

Conclusion

Both visible and invisible flicker from LED lighting causes a variety of adverse health effects and safety risks. Many building operators as well as affected individuals are unaware that flickering lights and video projectors may be causing these issues. Given the lack of legislation regarding flicker, the only way to eliminate the associated health and safety risks is to proactively identify and replace flickering light fixtures and video projectors and to avoid future purchases of these undesired products. Raising awareness about this issue will ultimately result in a healthier and more productive society.

References