Conducted RF

Discussion in 'The Projects Forum' started by rocky79, Apr 23, 2010.

  1. rocky79

    Thread Starter New Member

    Apr 23, 2010
    6
    0
    Hello,

    I am designing an LED board that is driven by Switching regulator(Buck) the Buck regulator frequency will be running anywhere between 200~ 300 Khz.

    Now the board has to pass the conducted RF interference test and conform to the chart below.

    Is it good idea to use Pi filter at the input ? What's a good design practice and how do i go about designing this filter so that i comply with the requirement in the chart?

    Thank you
     
  2. Papabravo

    Expert

    Feb 24, 2006
    10,144
    1,791
    In order to know how to interpret the chart you need to know how the test is conducted. That would be a good starting point.

    The units on the vertical axis are dBuA which I interpret as dB with respect to an RF reference current of 1 uA. Now 53 dB with respect to 1 uA is:
    Code ( (Unknown Language)):
    1.  
    2. 53 = 20 log(I/1uA)
    3.  
    4. 53 / 20 = log(I/1uA)
    5.  
    6. 10^(53/20) = I/1uA
    7.  
    8. 446.7 uA = I
    9.  
    Not sure what that level of RF current will do to a switching regulator, but it will sure make the S-meter on my receiver head for the pin

    Good luck in your quest
     
  3. rocky79

    Thread Starter New Member

    Apr 23, 2010
    6
    0
    Thank you for the insight Papabravo. One way to know if the circuit complies is by going to an EMI lab and experiment. But I was hoping to get a guidance on how to select a filter that gets me close to the specs.
     
  4. Papabravo

    Expert

    Feb 24, 2006
    10,144
    1,791
    What I'm not sure about is the precise location for a filter. On our high end audio equipment we use a common mode chokes and shunt capacitors for the DC supply coming onto a board. Again without being able to look at a complete schematic it is really hard to advise you.

    The common mode choke will pass low frequencies and the capacitor will shunt high frequencies to ground
     
  5. nanovate

    Distinguished Member

    May 7, 2007
    665
    1
    More important than the filter is the Buck regulator design and layout. The Buck topology draws AC current and causes a lot of conducted noise. If possible try to keep your duty cycle away from 50% as that is the worst case and operate in CCM. Use shielded inductors and minimize current loop areas. Once you have optimized that then you look at the filter. A properly designed pi filter or 2-stage LC (LC+ 6L4C) is pretty common.

    The RTCA DO-160 can be a rough test to pass. The set-up is usually a LISN and high-pass filter into a SpecAn
     
Loading...