Can transformerless power supplies be UL/CSA certified and is there any specific standard they must meet?

Question

I am trying to bias a linear optocoupler on the AC mains side of a circuit, in order to transfer analog information to the secondary side of the circuit. I have designed the following capacitive power supply to provide the 6mA required and it works.

However, the reviewer at the certification lab is indicating that the voltage specifications for the capacitors after the bridge rectifier must be rated for line voltage (~400V). I am skeptical of this, because this would make the capacitors massive to meet the capacitance needs and transformerless power supplies are typically smaller units limited by the size of the X capacitor used to provide the reactance. It also doesn't make sense, because the zener would burn out long before the capacitors could fail.

Even this application note suggests that the capacitors only need to be 2x the zener voltage:

Microchip Application Note AN954 - Transformerless Power Supplies: Resistive and Capacitive

Does anyone know the specification that would govern this or have appropriate experience in getting a power supply like this certified, to let me know what to look for? Can these power supplies be certified or are they an internet relic?

Thanks.

Answer 1

Actually, isolation is not a must for UL certification. So yes, a transformerless regulator can be certified. You can find a lot of non-isolated UL-listed products on the market.

Probably the reviewer has no knowledge about capacitive power supplies but anyways it seems that s/he treats C2 as a "Safety-Critical Component" due to its position (i.e. mains side - after the bridge rectifier). S/he may also suggest you to use even an X/Y cap there.

In one (or some) of the EN-.... standards (I cannot really remember, sorry) safety-critical components should be defined as:

• Input connectors
• Fuses & fuse holders
• Filtering components (capacitors, chokes etc)
• Varistors

etc.

From my point of view, the safety-critical components are R1, R2 and C1 (and a varistor, if any) in your circuit even if it has a zener at its output.

I think it's hard to find a standard document (because nearly all of them are protected against public access) but you can persuade the reviewer about the voltage rating about C2 by showing some measurements, test results, technical documents etc.

Answer 2

I don't have particular experience with transformerless supplies but have gone through safety testing numerous times, so I can offer some general advice.

One thing to keep in mind with regulatory testing is the concept of a 'single-point failure'. The safety reviewer might be considering what happens if C1 or R1 fails short. I suspect the voltage after the bridge might end up higher than you expect in this case.

One possible way around this could be to put two capacitors in series, or to make C1 a UL-recognized capacitor (with the backwards RU symbol on it). Recognized components have a level of 'trust' to them when it comes to single-point failures, and putting two in series ensures that you still have some impedance if one capacitor shorts somehow.

R1 and R2 might need to be split into two series elements as well for the same reason (single-point failure). Are they "flame-proof"/fusible resistors? These types resistors tend to fail open if there's an abnormal and may also be acceptable.

Another thing which may help is a large crowbar circuit (like an SCR) after the bridge which would clamp any excessive voltage if there's a primary-side failure in the attenuation circuitry.

Keep an open dialog with your reviewer. Once you identify their concerns you will have a much better chance of getting your design certified.