SMPS Insulation Resistance Problems

[VanillaIceCream]

Hello.
So I have a 12V Switch mode power supply without a ground connection,
 I measured the resistance between the output and input pins and my insulation tester says its around 3.3GigaOhms (3300MegaOhms) at 250V DC

Does this mean my smps is sufficiently insulated to be used as an isolated smps?

And what even is the minimum insulation resistance for a smps to call it "isolated"?
I have 230V mains voltage where I live.

Help would be appreciated.

[bdunham7]

3.3G is pretty isolated, really, but you also would want to test it at higher voltages.  However, the coupling between the primary and secondary is more likely to be capacitive than resistive, so you really should measure the AC leakage current.

[Siwastaja]

Practical answer, if it's coming from any of the proper big manufacturers (i.e., not a shady Ebay thing) and it has the proper approval markings (like self-certification CE in EU), consider it properly isolated.

If it is an Aliexpress special, consider it improperly isolated even if it shows fine on the meter because transformer insulations for example can break down later if poorly done.

Insulation resistance tester is only one step in verifying the safety of the product on the production line and sometimes afterwards, but alone it is meaningless to assess the safety of the isolation.

Answer to your exact question would be to first find out under which jurisdiction you are then look at the relevant standards.

[penfold]

I measured the resistance between the output and input pins and my insulation tester says its around 3.3GigaOhms (3300MegaOhms) at 250V DC

Could you share any details of the insulation tester? As far as measuring a high resistance, any capable instrument would be fine, but it's not the most effective way of assessing the standard of isolation. The more conventional method would be to use a "breakdown tester" or hi-pot tester which maintains the high potential for a minute or more and measures static leakage (basically insulation resistance) but is also sensitive to the more sporadic partial discharges and things that aren't often picked up by "resistance meters".

In proving that a PSU output is isolated, depending on safety requirement, you would also aim to do a type-test on a few devices, essentially tested to destruction. The production-line hi-pot test voltage would be a relatively small fraction of this ultimate breakdown and not in itself really prove that anything is isolated, it's more about proving there were no glaring production errors and to give some confidence that it will meet the same spec as the type-tested unit. Also, note that any hi-pot test is actually partially destructive and you should be wary of repeating tests (within successively decreasing the test voltage) on a device you intend to use (though, tests with a megger up to 1kV would be fine to repeat).

There could also be some EMC suppression caps (Y-caps) in there whose impedance won't show up with a DC resistance test but could cause more leakage than you expect.

For what purpose are you wanting it to be isolated? For human safety? (and would the aforementioned human be wanting to be in contact with the output terminals for an extended period of time?) or something more benign like a sensitive measurement system that needs isolation?

[Faringdon]

Yes, you should put 3.5kV AC from pri to sec and , less than a certain current should flow...i think its a mA or so, over one second.

The standard isolation test for production is to put 500V AC from pri to sec, for one second, and m,easure the current, it shoudl  be less than 250uA.

As Penfold implies,. you never do the 3.5KV AC test on a unit that will be sold after production, because it would be weakened by that test. The unit that gets that test, will get thrown away.
To test for general safety, you need to be at least mains peak in your DC voltage........but even thats not enough, because the mains suffers transients up to 10kV and beyond, so your supply must offer some safety from that.....is there such thing as a totally safe SMPS outh there, that would be safe against loads of repeated 6kV transients?......probably not, they all would eventually break down and become unsafe, but i guess there is always this trade off beteeen safety and cost.

I mean, your protection against a tranisnet is usually a MOV......and they break down and die after a certain number of transients.

[G7PSK]

Most SMPS will cease to function after a few HV transients long before the let the HV through or at least that is my experiance.

[Doctorandus_P]

I fully agree with Siwastaja.
If it does not have the proper stamps then it's not fit for mains voltage.
There is unfortunately too much shady stuff sold for low prices these days.