Shield Winding
Posted by: koji
on
Hello,
I am designing a power supply circuit using INN3870C.
I've built some prototypes and am running into EMI issues.
When measuring radiated noise, there are large individual differences in the noise level around 140MHz to 190MHz.
I think the reason for the variation is that there is a large variation in the way the transformer's shield winding is wound.
How should I wind the shield winding?
Is it better for the bobbin to be flat from end to end without any gaps?
Do you have any other comments?
How do you decide the number of turns?
please tell me.
Best regard.
Thank you for your interest in our products. One way to reduce the variation in the shield windings is to fill the entire width of the bobbin with the shield layer.
When the entire width is filled with the winding, there is not much room for the position of each turn to change between different transformer builds. In order to fill the entire bobbin, you can experiment with different number of strands with small diameter wires (e.g. AWG 30-35) to fill the entire width.
As for the number of turns, the easiest way is actually to just create multiple transformers with different number of shield turns. The interactions in the power supply that are important in EMI are extremely difficult to model so getting the correct number of turns analytically is a really hard task to take. Be aware though that changing the number of strands to fill the winding window WILL affect your EMI so that is another variable to consider when optimizing the number of shield turns.