For a "mostly" shielded sequence of 1 W where a user will see light bleed through small gaps in a device and (for argument's sake) might see a direct reflection of one of these 1 W LEDs (a multi-die LED with red and green sources), are there any obvious national/international regulations I'm going to need to be in compliance with that relate specifically to optical brightness? Not EMI/EMC restrictions, not electrical compliance, just optical brightness, like for lasers with e.g. 3A/B classifications.

  • 2
    \$\begingroup\$ ANSI z136.1 covers exposure of high powered light sources to the retina. You should make sure you are under the exposure limits defined in the spec, which is usually not too hard with LEDs unless you're relatively close to them. \$\endgroup\$ Commented Feb 27, 2023 at 2:52
  • \$\begingroup\$ How far away are the people's eyes from your 1W leds? Are there any lenses or curved mirrors? \$\endgroup\$
    – jonathanjo
    Commented Feb 27, 2023 at 13:10
  • \$\begingroup\$ So why did you tag this EMC if that's not what the question is about? \$\endgroup\$
    – Lundin
    Commented Feb 28, 2023 at 11:11

1 Answer 1


Obviously, for any safety-related work, it's your responsibility to ensure you're complying with up-to-date legislation for the appropriate jurisdiction. The following is from experience in the UK installing laser systems. The following should be taken as guidance on general principles only.

The legislation on non-ionising radiation defines limits in terms of how much energy can reach which parts of the body. As the most sensitive part is the retina, and the highest (spatial) density energy is from lasers, we normally concentrate on those, but the same regulations cover other kinds of bright lights such as welding arcs, other parts of the body in particular skin. I believe the core documents are IEC and are substantially the same in most jurisdictions.

It boils down to this:

  • You work out the divergence of your light source. In a laser, all the energy is a in a small beam, with a divergence of a small angle, perhaps 1 mrad. For a wider beam, the energy is limited by the human pupil, nominally a 7 mm circle
  • The blink response is nominally 0.25 s, which limits the time of exposure
  • Your overall system will allow people to get only within a certain distance by physical limitations (perhaps it's mounted on the ceiling)
  • You calculate your irradiance (power density) from your optics and power and exposure distance. Result in W⋅m-2 And then the energy in J⋅m-2
  • You compare your energy with the maximum permissible exposure (for skin or eye as appropriate), which is typically calculated for a given length pulse with MPE = 18⋅ t0.75 J m-2 (eyes visible light, 1 ms to 10 s duration, from EN60825-1:1994 via HSG95.)

In practice you're using a reflected source with high divergence, so you're almost certain to be way within limits for the light exposure from the brightness. But you still have to do the calculations, obviously.

You're still going to have to address risk factors other than specifically brightness per your question. (For example, in ships there are regulations about disturbing the eye's night-adjustment for devices on the bridge.)

A common other issue for regulation is for stroboconvulsive triggering, which is normally done by keeping strobing below 5 Hz.


  • AOR A very good starting point is the guidance on "artificial optical radiation" from the Health and Safety Executive, a British government body. link.
  • CAORAWW The Control of Artificial Optical Radiation at Work Regulations 2010 link
  • HSG95 A good summary is in in The radiation safety of lasers used for display purposes (HSG95) from the Health and Safety Executive, copy.
  • EN60825-1 European standard Safety of laser products - Part 1: Equipment classification, requirements and user’s guide The IEC and the BSI have matching numbers for the same standard. Superseded 2007 version
  • Changes in IEC 60825-1 The new edition of the international laser product safety standard IEC 60825-1 link
  • \$\begingroup\$ As a simple rule of thumb, would the laser safety thresholds be a good guideline for when you need to run the calculations? It other words, if your LED, treated as laser, would be considered safe under the laser safety regulations, then you should be fine, whilst if it is higher power than that you should run the calculations to ensure its safe? Because it seems like a laser is the worst case for divergence. \$\endgroup\$ Commented Feb 27, 2023 at 12:54
  • \$\begingroup\$ @user1937198 yes this is absolutely correct, you can use lasers as a worst case and consider yourself safe by a large margin. The only difficulty with this is the margin is enormous. For example, a projector with image of 1 m<sup>2</sup> is 25,000 pupil-areas, which dilutes the power a lot. Typically, too, non-focussed systems have such high divergence that once you get even a metre or two they no longer count as "bright". \$\endgroup\$
    – jonathanjo
    Commented Feb 27, 2023 at 13:06
  • \$\begingroup\$ Therese probably a large margin, but I was more thinking of lasers as the point at which you are safe enough you don't need to think about it, vs as you said, you need to run the calculations and work out appropriate precautions. After all, your projector could still be dangerous if someone is stupid and decides to sit directly in front of the lens and stare into it. Or are there some tables somewhere of equivalent power levels for 'so long as you stay 1m away whilst its active you are fine'? \$\endgroup\$ Commented Feb 27, 2023 at 13:14
  • \$\begingroup\$ @user1937198 For the most part, safety specs define permitted energy per angle/area onto the retina without specifically defining the generation mechanism. So if X is safe generated by a gas laser, it's also safe generated by a laser diode or a diode that is emitting without lasing. In practice though lasers are much brighter so it's usually a laser that, after calculating, you find is dangerous. \$\endgroup\$ Commented Feb 27, 2023 at 13:40
  • \$\begingroup\$ @user1937198 I use 1 metre square at 1 metre distance for my benchmarking (approx 1 steradian), and either blink-time or indefinite. For indefinite, the limiting case is a power 0.01 W/m2 (in HSG95 and IEC), and as that is 25984 retina areas, it is is 260 W. By normal standards this is very bright indeed! Full overhead sun is approx 1000 W/m2. All of these are too bright for sustained viewing But that shows you how bright, or how localised, you need to be to violate the intensity exposure MPE. \$\endgroup\$
    – jonathanjo
    Commented Feb 27, 2023 at 13:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.