How do AC and DC hipot tests compare?

Suppose I have a DC/DC converter, Recom RPP30 for an example to hand. It lists an isolation voltage of 3000 VDC. How does this translate to an AC isolation voltage? Would this be equivalent to 3000 VAC RMS? 2121 VAC RMS (3000 peak)? Something else?

• Sorry you are getting spammed, just protected your question to help. – Kortuk Sep 10 '13 at 14:24

Using the example you have indicated i.e. the Recom RPP30, it is certified UL 60950-1 (PP21) and this makes it a little bit more special than just any old dc convertor with isolation between input and output.

The baseline for a dc-to-dc convertor is that it should isolate but, because it is powered from DC, there may be no expectation for such a basic device to offer full reinforced insulation. This is because it's likely to be fed from an already "safe" ac source from a transformer and rectifier; the transformer (in most applications) will offer the safety insulation features.

However this one is approved/certified to 60950 and (as far as I can tell) it should withstand twice rated voltage plus 1000V. For 250Vac this works out at 1500Vac. I guess the assumption here by Recom is that it could be powered from a dc supply that derives its dc voltage directly from the AC mains power.

How does this tie in with what it says it is tested at?

They say it is tested for 1 minute at 2250Vdc and I guess this is close enough to 1500Vac (pk) so that it doesn't matter.

$1500 \times \sqrt{2} = 2121volts$

Maybe they decided to add an extra 129 volts on for good measure; hopefully someone can offer a better reason? Maybe they designed it to be safe when directly connected to an ac supply of 296V? This would become: -

$\sqrt{2} \times (2 \times 296 + 1000) = 2251volts$

It's interesting to note that 3000V is precisely one-third more than 2250.