Transformer Isolation Voltage: VAC vs Vrms

Power Transformers
, ,
1

Hello,

I have recently designed a power supply circuit using the flyback converter topology to step down my input voltage and produce a 5v output supply to power an embedded systems project I am working on.

One key design consideration was meeting a requirement for 2kV galvanic isolation between the input and output stages of this converter. When choosing a transformer I selected [THIS - 750319924] model which is available through DigiKey.

At the time I selected it I was operating under the assumption that I could calculate the peak voltage isolation by simply multiplying the value given by the component page on DigiKey’s site (1500 Vrms) by sqrt(2) which gives 2115 Vpeak.

However, after a colleague pointed out that the official datasheet for said transformer lists the isolation test voltage as 1500 V (AC).

Now I have confused myself and I am unsure if this is the same thing. A cursory google search tells me that V (AC) =/= Vrms. Can anyone provide more information or clarify this for me? I am wondering if the product page on DigiKey incorrectly listed the component’s isolation at 1500 Vrms when it should actually be 1500 V (AC). Are these really the same thing?

I am stressing about whether I need to change out my transformer for something with a higher isolation rating, but doing so will increase cost and require a larger physical footprint on the final pcb.

Any assistance would be greatly appreciated. Thanks!

2

@patjwbrown

I find the same information for the item you had ordered and swapping it our for one rated for your needs would be needed.

Looking over options that are otherwise similar I find the below.

750313109
Data-sheet shows a value of 2400VDC (~3394vac)

750313109 Würth Elektronik | Transformers | DigiKey

PDC023-EP13A05S
Data-sheet shows a value of 3000Vac

The sizes for these are different however they are not super huge either.

750319924 = 13.34mm x 11.43mm tall x 15.24mm
750313109 = 9.78mm x 10.54mm tall x 9.5mm
PDC023-EP13A05S = 17.5mm x 13.5mm x 13.00mm tall

I would have gone with the PDC023-EP13A05S but it is not an item we keep in-stock.

The 750313109 ($7.68USD) is a similar price to the 750319924 ($8.24USD)

As for the data being listed as it is on the site I have submitted an inquiry to have it reviewed.

3

RMS values indicate the equivalent-power DC value of a time-variant source. For a sinusoidal waveform, the peak value is indeed ~1.4x the RMS value.

That’s somewhat irrelevant in questions of insulation ratings and specifications however. Like most “how much before it breaks?” questions, a specific value for a given device cannot be established without destroying the part.

People work around that by specifying proof standards, wherein a specified stress in excess of normal working values is applied, in the understanding that if a device is shown to withstand it, then it can be reasonably expected to withstand something less.

It would be prudent to unwrap the specifics of that specification. To withstand a peak value for, say, 5 seconds is one thing, to hold up under the same stress for 30 years, quite another.

What the datasheet value for the part in question indicates is that an AC waveform with a 1500V RMS value was applied between the windings indicated for some period of time, and the resulting current flow was not in excess of some value that would indicate an insulation failure.

In terms of your product meeting the requirements that apply to it, that information is useful but insufficient, because there’s nothing stopping you from putting other components of lesser fortitude across the isolation barrier, violating creepage/clearance norms, or committing any of several other sins that may potentially result in weeping and gnashing of teeth.

At day’s end, depending on how your specification is written you’ll need to perform your own high potential (HIPOT) test of the completed assembly to prove that the aggregate device is indeed capable of withstanding that stress. Choosing a transformer that’s already been proven to that same level is a really good option if one’s interested in a good production yield and limiting the chances of somebody else questioning one’s judgement at a later, post fit-meets-shan point in time.

Insulation ratings and related standards are a topic that one ought to be informed of if one’s going to dabble in areas where it’s relevant. To that end I found this FAQ from ADI that dangles the ends of several important conceptual threads out in the air for a person to grab on to. It’s written in context of their digital isolator products, but the ideas are highly transferrable to other situations where the same concerns arise.

P.S. Have you given any thought to EMI compliance…?

1 Like