What is safe isolation resistance for mains connected LED product?

Hello,
We have some non isolated, offline (240VAC) mains 4W LED lamps in a tube. (so its like a mini fluoresent tube but has LEDs in it) The PCB is glued down to a metal heatsink which runs along the back of the PCB. Before shipping, we do an isolation test whereby we tie Live and neutral together, then put 500VDC between the heatsink and the Live/Neutral.

Anyway, some of the units have been blowing up in the field. When the units have blown, they have some blackening in the Perspex tube, but none of the components look damaged, and they still work fine.

When we repeat the hipot test on the blown tube, it reports a resistance of 130 megaohms between live/neutral and heatsink, whereas a newly assembled unit reports open circuit (too high resistance for the hipot test kit to be able to measure, but its a pass so we accept that).

So my question is, is 130megohms of resistance between live/neutral and heatsink acceptable and within the regulatory limits? Or is it classed as being unsafe?

130 Mohm is no problem as such, but the change might indicate a problem. 500V hipot test is inappropriate for mains isolation, review IEC1010 or other applicable standards for required test voltages.

130MΩ Insulation resistance between line+ neutral to heat sink is quite reasonable. As per UL standard 100MΩ minimum insulation required. Normally the IR meter shown resistance in GΩ and TΩ if there is enough space. IR values are not consistent it will vary with humidity. If your part having 130MΩ IR it means, there is no arcing up to 500V or slightly above voltage.

The measured resistance at 500V isn't relevant. You should be testing with at least 1.5kVAC.

thansk, but these tubes are then installed in a kind of enclosure, so people cant actually get their fingers on the heatsink when its in the enclosure. So i believe 500V is fine for the isolation test.

First you are asking for safe isolation, then you say you don't need isolation at all?

i am certain there is a mains 500v isolation voltage spec'd by NETA or something?

I had mention in #3 100MΩ minimum insulation resistance required for UL/CSA certification but for ANSI/NETA standard, 5MΩ minimum insulation resistance required for 250V appliance. UL standard is more concern about human safety.

If I remember properly, per UL the hipot voltage should be 1000 v plus twice the line voltage.
In your case, 1500 v.

If I remember properly, per UL the hipot voltage should be 1000 v plus twice the line voltage.
In your case, 1500 v.

Click to expand...

Thanks, but there is also an insulation test as follows (attached).
Thsi was done by our chinese smps design/manufacturer for our 60w offline flyback.
As you see its done at 500V

This was done by our chinese smps design/manufacturer for our 60w offline flyback.
As you see its done at 500V

Click to expand...

Proves what particularly? Surely not that the power supply complies with a specific safety standard.