1 minute vs 1 second isolation voltage test for isolated offline SMPS transformer

[treez]

Some offline transformers are spec'd for 3750vac for 1second (3mA).
Some are spec'd for 3750vac for 1 minute (3mA).

Generally, which applications require 1 minute?

..is it mostly industrial apps?

 

[wwfeldman]

the transformer tested for a longer time is likely built better

i expect more critical application require a longer isolation test

what is the 3mA about?

 

[treez]

The 3mA is the limit value for the leakage current that may flow when under the 3750vac stress for the given time....any more than 3mA flows and its a fail.

 

[FvM]

Most safety standards, e.g. IEC 1010 are requiring 1 min test, I expect that it must be used for type approval. Routine tests in production will probably use shorter test time.

I wonder which standard is requiring respectively allowing 1 sec test?

 

[wwfeldman]

this test is accomplished by tieing all of the primary leads together and all of the secondary(ies) leads together
and then applying 3750 VAC

3 mA seems like a lot of leakage

 

[Easy peasy]

the 3mA is capacitive current - not necessarily leakage -

usually the test is 2750 for 1 min or 3750 for 1 sec

& 3750 for 1 min or 5000 for one second

depending on the spec being tested to

 

[treez]

Thanks.
It is strange to me that power integrations, on their design software, only ever design the isolated flyback transformers to 3000vac for 1 second (thats for 200-265vac input).
(screenshot of design result attached)
Do you know why this is?

Why would anyone want to spec it for one minute?......no mains transient lasts for one minute?....they only last 50us

 

[c_mitra]

Some offline transformers are spec'd for 3750vac for 1second (3mA).

You forgot to mention the frequency.

This is basically a test for the dielectric. Some certification agencies require the isolation test.

The voltage must be ramped slowly (else your test will fail).

If you pass the 1s test you will most likely pass the 1m test too (unless you have lots of losses in the dielectric).

You need to consult the standards for details (I have zero idea).

- - - Updated - - -

no mains transient lasts for one minute?....they only last 50us

If you apply 3750V for 50us, you will not pass the test. So better way is to slowly apply 3750V for 1s and you will be ok.

I am not sure about the 1m test; some dielectrics fail slowly.

 

[FvM]

I found that IEC 61558-1 (Safety of power transformers) requires 1 min high voltage test for compliance tests and 1 sec for production tests. There may be additional requirements in IEC 61558-2-16 (Particular requirements and tests for switch mode power supply units and transformers for switch mode power supply units).

You forgot to mention the frequency.

It's 56/60 Hz for AC tests.

 

[Easy peasy]

The whole reason for the 1 second test - at the higher voltage - is speed of manufacture, over the 1 min test. The voltage is scaled up ( not linearly ) such that the 5kV ac for 1 sec is equivalent to 3750Vac for 1 min - is equivalent to 1kV ac for 5 years ....

 

[treez]

Thanks very much..

We are wondering what kind of isolation test the following power supply is spec’d to?…..
It’s the Philips Xitanium Xtreme 75W LED driver…..AKA…
“Xitanium 75W 0.35-0.70A GL Prog sXt” (datasheet below)
The datasheet says it has “Basic” insulation.

That is, how would they have carried out the hi-pot test on it? (voltage setting?, time duration?, current limit setting?)

The following reveals that “basic insulation” means withstanding 1000V+2x(operating voltage)…however, I happen to know that Basic Insulation means 3750VAC for one minute for an isolated offline SMPS…..during that minute, the RMS current should not exceed 3.5mA.
https://electrical-engineering-portal.com/what-is-hipot-testing-dielectric-strength-test


Would you agree that at the Philips Approvals testing session for Dielectric withstand for this power supply…they would have set the Hi-pot instrument to 3750VAC for one minute and set a current limit threshold of 3.5mA?

(the standards do not actually specify a current limit threshold)


Philips Xitanium Xtreme 75W LED driver datasheet:
https://www.docs.lighting.philips.c...m_75W_0.35-0.70A_GL_Prog_sXt_929000702302.pdf

Here is the product brochure
https://www.docs.lighting.philips.com/en_gb/oem/download/xitanium/Xitanium-XtremeLED-brochure.pdf

 

[Easy peasy]

They list the standards in the data sheet - it is a matter of reading same.

 

[treez]

Thanks, but we notice that IEC61558 isnt mentioned in the datasheet

 

[Easy peasy]

IEC61558 has been withdrawn - it is no longer an applicable standard - from my recent information received.

 

[FvM]

At least EN 61558 is still applicable in Europe, the UK standard has been recently released: BS EN IEC 61558-1:2019-06-25

If it has been withdrawn in other domains, what replaces it?

 

[treez]

thanks, all we want to do is repeat the same hi-pot test that Philips did for their approvals testing, to be sure that the power supply has sufficient isolation. But from the datasheet, there is no way of knowing what hi-pot test they did. I have written to Philips before for technical info….all you get is a load of emails asking what your predicted sales volumes are, then they never get back, even if you tell them you've got millions of sales lined up.

 

[FvM]

You have previously written that the power supply is classified as basic insulation. Then the test is surely not performed with 3750 VAC. EN 61558 is e.g. specifying 2100 VAC for 300V working voltage and basic insulation.

I can't tell you what's the applicable standard for LED power supplies, because it's not my business. I presume that 61558-1 and 61558-2-16 are not completely off topic.

 

[treez]

Thanks, is their any way for us to know how they did the hi-pot test on the mains transformer in the power supply?
Ie what voltage time and current setting for the hi-pot test?

We wish to remove one and do the test on it before buying a big batch of these power supplies.