• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 61 subscribers
  • Views 804 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.

TPS92075 EMI filter

Oleksiy Pazyuk
Expert 2585 points

Other Parts Discussed in Thread: TPS92075

Hi,

What is difference in using different kind of EMI filter for  TPS92075?

I see that for reference design PMP6009    non-symetric pi-filter is used, while for evaluation module (TPS92075EVM) t-filter with snubber is used.

Is there any suggestion for  calculation of such EMI filters?

  • 0
    TI__Intellectual 1115 points

    I'll try to answer as simply as possible, the complete answer requires a good amount of power electronics experience and triac dimmer compatibility knowledge to fully understand.

    i) PMP6009 use a 2nd order filter , other designs use a 4th order filter. In general a 4th order filter gives better triac compatibility then a 2nd order because a 4th order allows the use of less capacitance. However, depending on the power level a 2nd order may be fine. A second order uses less components, but they must be larger in value.

    ii)You can get your attenuation using either method or any one of several other filter configurations, in the end you need a filter to provide Xdb of attenuation to pass conducted levels.

    iii) To calculate the required attenuation you need to start with the waveform shape of the main switch current, which is the main source of all the conducted noise.  Use some analysis software to perform a fourier analysis on that waveform to obtain the fourier components.

    iv) Take the fourier data (output from a .FOUR command, for example in spice) and convert it to a dBuV noise level. 

    v) Use the highest harmonic that falls under the conducted EMI frequency range of 150kHz to 30MHz, then find the QP limit for that frequency from the standard. For example, if your analysis showed a fourier component of .0953 at 800kHz, that would represent a noise level of 133dBuV.

    vi) The QP limit at 800kHz is 56dBuV. So we would need ~77dBuV to be under the limit. We would then design the filter to obtain >77dB of attenuation at 800kHz

    v) You need compute the filter attenuation by hand or use some analysis software and adjust values accordingly.

    vi) When plotting your frequency response you can also examine the filter break points and add components for damping.

    OR - Alternate Method:

    i) Start with one of the filters provided on various EVM's and Reference designs that are close to the power level you will be running at.

    ii) Try it and modify components until your design passes conducted EMI levels. Larger values of L and C = more attenuation. :-)

    Thank-you for your very interesting question.

  • 0 in reply to Tim Sullivan
    Expert 2585 points

    Hi Tim,

    Thanks for your comprehensive answer.

    But my concern was about asymmetry  in the EMI filter design. For example  in the PMP6009   I see  that in  the filter is used  different values of  resistors R7 and  R1.

    This slightly confuse me. Because if we talk about 2-nd order filter I suppose that branches at the both side of the inductor should be the same. Is there some sense in such asymmetry ?