1
\$\begingroup\$

This question is a follow-up to Does switched-mode power supply feature galvanic separation? : Over there, stevenvh says:

"... a mains connected SMPS, and most of those do have galvanic separation."

How do I find out whether a given SMPS (say, my cell phone charger or a similar device, so something without a datasheet readily available) features galvanic isolation?

  • If there is no third (earth) connection on the mains plug, is that enough to conclude that the DC output is separated from mains, i.e. do regulations like CE require isolation for non-earthed devices?
  • If that's not enough - is it sufficient to test both plaugs on the AC side against both contacts at the DC side with a multimeter, or are there designs that are not separated despite having no connection when turned off?
\$\endgroup\$
4
\$\begingroup\$

With a multimeter. Measure the resistance between various "output" pins and the "input" pins.

A very high resistance (>1 meg ohm) or an open circuit means that it is isolated. A resistance of less than 10K ohms indicates that it is not isolated. A resistance in between those two values means that something is weird.

\$\endgroup\$
2
\$\begingroup\$

If you really want to be sure you need to do Hi pot testing: -

enter image description here

Here is a great document from XP that explains it. Class II is what you are looking for i.e. no earth connection. You cannot check that you have galvanic isolation with a multimeter and realistically hope to believe it although you can definitely disprove galvanic isolation if you find some degree of conductivity.

\$\endgroup\$
7
  • \$\begingroup\$ You cannot verify that the supply meets regulatory safety requirements with a multimeter. But you can certainly see if the supply is isolated. \$\endgroup\$
    – user3624
    Jun 10 '13 at 21:25
  • \$\begingroup\$ @DavidKessner If it were the audio power output on a PA feeding several distributed loudspeakers I agree that the multimeter test is sufficient to prove the PA uses an output transformer but not on a SMPS. I guess I'm saying that galvanic isolation doesn't necessarily mean safe. \$\endgroup\$
    – Andy aka
    Jun 10 '13 at 21:55
  • \$\begingroup\$ You're right that Galvanic isolation does not mean safe. But assuming that the power supply was designed correctly and passes the appropriate regulatory tests (which would be the case with a commercially bought supply) then a multimeter test should correctly indicate if it is isolated or not. As would reading the datasheets, if available. To be absolutely sure, you need to do a Hipot testing, as you mentioned, but few people on EE.SE have the equipment or knowledge to do that. \$\endgroup\$
    – user3624
    Jun 10 '13 at 22:24
  • 3
    \$\begingroup\$ Do you see a case where a commercially available and undamaged supply would say "isolated" according the multimeter test, but not actually be isolated? I don't, but maybe I am missing something. \$\endgroup\$
    – user3624
    Jun 10 '13 at 22:26
  • 1
    \$\begingroup\$ I can imagine a circuit for a non-galvanically isolated SMPS that would need the meter ohms measurement to overcome 2 x diode drops to register a reading. It's hypothetical of course. Input bridge rectifier and output bridge rectifier with secondary winding connected to neutral. I can't imagine such a thing exists. I guess anything with diodes/semis in could fool the meter if it couldn't overcome a certain forward voltage. \$\endgroup\$
    – Andy aka
    Jun 11 '13 at 14:02

Not the answer you're looking for? Browse other questions tagged or ask your own question.