• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 592 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CM balance technique in flyback converter to reduce CM current

Remigiusz Buszka
Prodigy 20 points

Hello

I would like design flyback transformer  (about 10W)  and reduce CM current using CM balance technique.

In TI Literature Number: SLUP338  from Power Supply Design Seminar 2016 page 24 describes this technique.  ( by Isaac Cohen and Bernard Keogh )

www.ti.com/.../slup338.pdf

I have  windings

core

Primary 30T

BIAS  10T

SECONDARY 8T

Shield 50um copper foil

Pimary 30T

I know that I should connect shield to tap of BIAS , Naux=1/2NS = 4T   (similar to figure 33) but....

in SLUP338 is another requirement,  Nb1+Nb2=Ns    (number turns of bias=number turns of secondary)

My  Bias voltage must be > secondary voltage  and is a problem.. :-)

What should I do  to fulfill thic condition ? (Nb1+Nb2=Ns)

If Bias <Secondary , there is no problem beacause Nb2 my be not connected like in example on figure 33, but  if Bias >Secondary  ?

Should tap secondary after 8T  -> to output diode and add unconected 2 turns ?

Please help

Thanks!

  • 0
    TI__Guru* 85616 points

    Hello,

    This technique only works when the aux winding turns are less than or equal to the secondary.

    Is there any reason why you can't make your aux winding 8 turns instead of 10?  This removes the need for Nb2 if you do it.

    Regards,

  • 0
    TI__Genius 9130 points

    Remigiusz,

     

    There are always many ways to tackle CM balance, and the various examples in the papers are just that - examples. Since every design is a little different, the CM balance solution will need to be adapted to suit each design. And very often it will need to be modified and iterated a few times to try to get the best result.

     

    In this case, since Nb = 10T but Ns = 8T, there are a few possibilities.

     

    The 2T imbalance between bias and secondary could be addressed by winding a 12-T bias, but winding the last 2T in the reverse direction, in anti-phase with the previous 10T. That results in a net 8T effective on the bias layer, which will balance the interface between bias-sec. But the entire 12-T bias layer must be a full single layer (using multiple thin parallel strands) to completely shield noise from the half-primary underneath. Clearly there will be practical challenges with implementing this kind of winding, and the multiple taps and reverse winding for the last 2T will somewhat compromise the shielding effectiveness.

     

    Another possibility is to wind the 10T bias layer as is (again in a full single layer that is completely full for shielding purposes), and then adjust the tap that drives the foil shield to maybe 3T instead of 4T – that way the current injected into the sec layer from the bias layer due to the 2T CM imbalance is partly compensated by a similar amount to CM current flow from sec layer into the foil shield. A degree of fine-tuning can be achieved by choosing the point along the shield length where the 3T tap is connected. Generally the shield is connected to the 3T tap in the middle of the shield. Since the shield is a single turn itself, with induced voltage, the effective CM signal on the shield will be ~+3.5T at one end, and ~+2.5T at the other end, for a net effective average of +3T by tapping in the middle of the shield. But it’s also possible to tap at one or other end of the shield instead of the middle, to get an increase or decrease in the average CM voltage by ~0.5T. Or it’s even possible to get a different fraction of a turn or higher/lower CM injection by choosing the point along the shield length to connect the 3T tap.

     

    There will also be other possibilities to get CM balance depending on the practicalities, but as I said, every design is unique and will often need its own appropriate CM balance solution.

     

    I hope this is helpful for you.

     

    Regards,

    Bernard

  • 0 in reply to Bernard Keogh
    Prodigy 20 points

    Thanks a lot for the explanation.

    I will try to wind up several types of transformers and check the level of conducted disturbances.

    I will write what were the results of each solution.

    For my part, the solution that came to my mind was as follows, wind an additional 2 windings on the secondary winding

    (not connected, in phase with the secondary one)

    Thanks again

    Best regards.

    Remigiusz Buszka

  • 0 in reply to Remigiusz Buszka
    TI__Genius 9130 points

    Remigiusz,

    Yes, your proposed solution can work too, and it's probably a bit more straight-forward to implement, and may give better CM balance performance. As I said above, there are always many alternative ways to achieve CM balance, it comes down the practicality of each and finding the solution that is most cost effective for each individual design.

    However, adding extra dummy sec turns does sacrifice sec winding performance - since the sec needs 25% extra turns, that will reduce the available winding width for the active 8T, forcing the use of smaller diameter wire and/or less parallel strands. That will increase the sec DCR/ACR and push up the copper loss. Sec copper loss can be significant, since the sec RMS current is usually quite high in a Flyback. There is also the complication of tapping at 8T from what is typically heavier guage wire (and/or multiple parallel strands), and the added interconnect can compromise subsequent winding layers.

    However, if that is acceptable in your case, this could be a good viable solution for you.

    Best of luck with your design.

    Bernard