UL / IEC Spacing vs Potting

Posted by: Imre A. Csaszar on 09/05/2008

Hi, I'm familiar with dielectric/arc and creepage distances of UL/IEC 1950. The design I'm considering is a universal input 25W supply and I intend to make it as physically small as possible, my question is is this: Can I eliminate the mandated distances by vacuum encapsulating the whole thing? After all, signal transformers, opto-isolators and power transformers usually have a 1mm or less separation between line and output conductors. If the entire assembly is potted with internal distances between parts & conductors so that the potting dielectric provides more than 5kV isolation, and the final supply passes all leakage arc over tests, do the safety agencies approve this? Thanks for your insight, Imre A. Csaszar P.S. I just finished a design using the LNK304D & PIXLS (5VDC @ 50mA non-isolated for a uC & signal LED). After reading the Tech Data, I am amazed how easy this is! Thanks for all the excellent, thorough, detailed, and practical info on your site. I started by looking at the Supertex SR087 at first, but the whole thing is kind of flakey and the Supertex App engineer recommend against using this part for ANY designs.

Comments

I am not a safety expert but I have seen designs where the transformer has been fully potted (encapsulated) to give smaller size. I do recall that the transformer had to be sectioned (cut into slices) to prove that there were no air voids in order to receive agency approval.

Hope that helps. Anyone out there have a more thorough answer?

Cheers

PI-Chekov

Hi Imre, may I ask you to show the small PCB you've created because I am struggling with an audible noise problem while minimizing the PCB using LNK304 and partially SMD parts. thanks, Zedman

Imre, You should start by purchasing the standard that will be used to evaluate your design. You mentioned 1950 (now superseded by IEC 60950 3rd edition). Assuming that this is a 60950 evaluation, I would argue that your vacuum encapsulation meets clause 2.10.8 and therefore only requirements for through insulation distances should be considered and creepage/clearances ignored. Clause 2.10.5.1 specifies what these distance are: 0.4mm for reinforced insulation. They may hold you to typical creepage/clearance requirements on the board if they think that tracking could occur on the board surface (and note in Table 2N that on a coated board the minimum separtion of outer conductors at 320V is 1.6mm). I would try to argue that if the board is entirely encapsulated it should be treated the same way that inner layers of printed circuit boards are treated (2.10.5.3) and that is distances through solid insulation (2.10.5.1) where the requirement is 0.4mm. The encapsulation will have to pass thermal cycling tests, dielectric strength tests and you will most likely have to do a routine production dielectric across the separation involved. Consider also doing a preliminary construction review with your agency. Mike

Hi, take into account that all the manufacturers are not able to get enough vacuum to avoid air bubbles (no matter what they say). I advise you to cut the firt sample and check that the transformer is solid as a rock...