Evolving Laser Safety Classification Concepts & New Products

By Karl Schulmeister

The classification of lasers by the product’s manufacturer – from Class 1 to Class 4 – is a valuable means to provide the end user with simplified information about the potential hazards to the eye and skin.

The concept of product classification can be considered a success story. Developed in the USA by the CDRH in the 1970s, it has been accepted internationally for more than 30 years, based on the standard IEC 60825-1. While the basic system of classification has remained unchanged since its inception, some adjustments were necessary over the years and will also be necessary for the future, when reacting to new types of lasers and scientific data on injury thresholds.


For a few years, diffractive optical elements (DOE) and microscanners have driven a large group of new products; mainly gesture controls and 3D cameras for consumer electronics (see Image 1), but also scanned lidars for machine vision and autonomous cars, as well as pico-projector scanners. For these new products, the combination of factors results in challenges for product safety and standardization. They are not intended as specialized professional products, such as lidars have been for the military, but are for consumer use. Therefore, in practice, they would need to be Class 1, Class 2 or Class 3R devices (depending on the wavelength range and country) but at the same time, for a satisfying performance in terms of detection distances, emission levels need to be relatively high. Because of the diverging or scanned nature of the emission, these systems suffer particularly from the conservative combination of classification rules of a 7-mm diameter pupil, an assumed exposure distance of 10 cm from the DOE or from the scanning mirror, together with an assumed accommodation to the apparent source at such short distance. While laser safety classification was always historically on the conservative side, it might be possible in the future to consider that the combination of those three exposure conditions is not only highly unlikely, but there are also reflexes (the near triad of accommodation) that result in pupil constriction when accommodating to a close target.

Defining measurement (pupil) diameters smaller than 7 mm for very close distances and as function of accommodation target might be a possible relaxation for future amendments, but would make the analysis even more complex. Also, possibly, emission limits can be raised somewhat in the higher nanosecond and lower microsecond regime, which is a task for the International Commission on Non-Ionizing Radiation Protection, ICNIRP to which the IEC refers for bio-effects committee work. Particularly for a change in the emission limits the general “predicament” exists that the injury thresholds depend in a very complex manner on wavelength, pulse duration and retinal spot size. When emission limits for products (or exposure limits for the eye) are to be made to reflect the thresholds more accurately to reduce needlessly large safety margins, it automatically makes the limits more complex since simple limits by default would be, for many scenarios, over-restrictive. One exception in the 2014 IEC and ANSI revision applied to small retinal sources, where it was possible to greatly simplify the analysis of pulsed emission by setting the multiple pulse correction factor CP (or C5) to unity, at the same time permitting significantly higher emission levels as compared to earlier editions. On the other hand, in the same revisions, the analysis of extended retinal images became more complex by permitting significantly higher emission levels for devices in the range of the lower “safe” classes.

Besides possible adjustments in the emission limits, two concepts based on engineering safety features are currently in development in the responsible standardization committee at IEC to permit higher emission levels for divergent or scanned systems – but still achieve classification as “safe” class, such as Class 1 for IR and Class 2 for visible emission.

The first is a virtual protective housing (VPH) where the emission is automatically reduced when an object enters the VPH. In such a device, one or more sensors monitor the protected volume. Outside of the protected volume, the emission needs to be below the limits for the class that is to be achieved, such as Class 1. When the VPH is free of relevant objects, the emission level within that volume can be higher: as long as human access to this radiation is prevented by the system, it is not relevant for product classification. The sensor system thus establishes a virtual protective housing instead of a real one, and defines what is referred to as the “closest point of human access”.

The second type of engineering measure to raise permitted emission levels applies to lasers mounted on vehicles and other moving platforms. When the vehicle is stationary, only normal emission levels are permitted. When the vehicle is at a certain speed, it can be assumed that another vehicle that is driving at the same speed will do so with a minimum distance. Thus the speed of the platform is the basis to define the closest point of human access that is to be considered for classification, which can, for instance, be 1 or 2 meters from the car with the laser.

Both types of engineering features have the advantage that the emission is tested against permitted levels at farther distances than usual, resulting in significant increases of the permitted emission level for diverging or scanned emission. While the IEC standard can already be interpreted in a way as to permit classification on engineering features that prevent human access, in order to assure international standardized testing conditions, it is necessary to update the IEC standard and provide specific performance requirements. For instance, for the virtual protective housing, it will be necessary to define probes used to test if the emission is reduced when an object enters the VPH. For the “moving platform” concept, it will be necessary to define the measurement distance as function of vehicle speed, as well as additional requirements to prevent that people on or in the vehicle have access to hazardous levels of laser radiation, such as when the laser is mounted on the roof of the car and there is a sunroof, or people on a pickup truck’s bed. A virtual protective housing might be needed to prevent access for these cases and to ensure that the concept of “moving platform” is internationally accepted for formal product classification. After all, it needs to be appreciated that classification of products following IEC 60825-1, as a basic principle, can only rely on engineering performance of the device and cannot depend on proper installation or behavior of the user.

**Several of the issues discussed in this article were also topics of ILSC 2017 papers, including the history of CDRH and IEC standards in invited presentations by Jerome Dennis and David Sliney, respectively, as well as the moving platform concept. The 2014 updates of IEC and ANSI standards were discussed in earlier ILSC papers.

Karl Schulmeister was project leader for the 3rd Edition of IEC 60825-1 and is a consultant on laser product safety at Seibersdorf Laboratories in Austria. For more information,
visit http://laser-led-lamp-safety.seibersdorf-laboratories.at.