In the days of old when Knights on horseback hunted dinosaurs by flashlight, electrical safety regulations were a much simpler affair. Incandescent light bulbs were installed in metal luminaires with a bit of Bakelite here and there for insulation. To make everything safe the electricity supply incorporated a fuse while an earth wire kept the metalwork at ground potential.
Then along came plastics, double insulation, residual current circuit breakers, fluorescent tubes, dimmers, and now switch mode power supplies and LEDs. To accommodate all these changes the regulations were expanded and improved.
Unfortunately, modern technology moves faster than regulation can keep up, and in the case of LED lights there is confusion as to what is appropriate and applicable.
From an electrical perspective LEDs are incredibly safe: They operate at only three volts DC, so you need a resistor to actually decrease the voltage before you can use an LED in a three cell flashlight. It also means that to operate most types of LED from the utility supply requires a power supply. These range enormously in sophistication. At one extreme it can comprise a bridge rectifier and with a capacitor used as a reactive dropper. These are common in the types of cheap LEDs bulbs you might commonly find online, where the phrase “caveat emptor” seems highly appropriate. What should be present is a properly designed switch mode constant current power supply with over voltage, over temperature, spike suppression, short detection and other protective circuits.
In an application like domestic lighting, white LEDs need to be in the 5 to 10 W power range to compete with incandescent bulbs on a Lumens basis. Because LEDs are relatively inefficient, albeit miles better than incandescent bulbs, they have to be soldered on special metal-in-board PCBs to dissipate the heat produced. The PCB connects to a heat sink, which will often be accessible as the fins need to be in fresh air to provide cooling. The question then arises: How much dielectric isolation does there need to be across the metal-in-board PCB between the finger-accessible metal side and the electrical tracks connecting the LEDs to the power supply?
There are roughly 50 standards potentially applicable to LED lighting and many are still in development. Not only is there potential uncertainty over which standards are applicable, but standards are also complex documents that are written in carefully constructed English. There is good reason for this – they are intended to cover all conceivable scenarios within a precisely defined scope. Yet to an LED engineer (having a casual flick through in search of design guidance) they sometimes might as well be written in Latin.
I asked a few people in the industry how much dielectric isolation they thought was required for metal-in board PCBs used in LEDs. Unsurprisingly, the answers ranged from “50V” through “it depends” to “over 5kV”.
Making the correct choice matters because both the breakdown voltage and the thermal resistance between the LED and the heat sink are proportional to the thickness of the dielectric in the metal-in-board PCB. Keeping LEDs cool is important for sustaining their hue, brightness and longevity and also to maximise efficiency. Using a Nanoceramic for the dielectric means this layer can be extremely thin to provide good cooling yet with sufficient dielectric potential to ensure the product will be electrically safe.
Because electrical safety is of huge importance interpretation of the standards, and from there the dielectric potential of the metal-in-board PCBs, should always be left to those that are experts in the field.