3
\$\begingroup\$

I am making a charger and I am measuring conducted and radiated EMC. The circuit is a flyback and the transformer in mention is a sandwich style transformer with 4 windings. Primary windings, secondary windings and two sets for supply for IC's. All the windings in the transformer are wound the same direction.

My question is why does the radiated and especially the conducted EMC measurements get a lot better when I simply just switch the start and stop of each of this woundings, hence they are all the same direction so no phase shifting in the picture.

\$\endgroup\$
1
  • \$\begingroup\$ Please show winding schedule / construction diagram of the transformer. (If it's an off-the-shelf part, you can always take one apart and see.) \$\endgroup\$ Oct 4, 2023 at 17:04

1 Answer 1

3
\$\begingroup\$

In most non-RF transformers and inductors, the windings form many layers on the core.

In a typical switching converter, one side of the inductor or transformer primary is connected to the switching node, and the other side connected to the input voltage (boost, flyback, forward converters), the output voltage (buck converters), or ground (inverting buck-boost converters). The switching node produces lots of noise simply because it's switching, but the input and output voltages and ground are all comparatively noise-free.

If the side of the inductor/transformer that's connected to the switch node is the side connected to the outermost layer of windings, you can expect the inductor/transformer to give off a good amount of noise. But if it's the innermost layer of windings, the outer windings (which, remember, are connected to a pretty quiet node, either input, output, or ground) will act to shield the noise, and you'll get much less noise coming off your magnetics.

This is why many inductors are marked with a line or dot, even though they're nominally unpolarized devices. The marking indicates the side connected to the innermost winding, so that's the side you should connect to the switch node.

While it may seem like this would only affect the radiated emissions (and indeed, it does only affect radiated emissions directly), bear in mind that the closest things around that can pick up those radiated emissions are the traces on the very circuit board that the inductor is on--that coupling can convert radiated emissions to conducted emissions. Whether this is indeed why you see mostly conducted emissions being affected or not, I don't have enough information to say.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ @tobalt I missed that. Added a bit about coupling to the board that would cause radiated emissions to be come conducted ones. \$\endgroup\$
    – Hearth
    Oct 5, 2023 at 15:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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