Switching regulator "noise" means?

Discussion in 'General Electronics Chat' started by s_mack, Nov 17, 2012.

  1. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Hi. Simple question for those that know. I've read that a disadvantage of using switching regulators (vs linear) is the "noise".

    Is it rf noise generated by the circuit (that may affect other devices within a certain range)? Or is it "dirty power" that comes out which needs to be cleaned up? Or both, or other?

    What I'm wanting to do is provide steady 12v to a camera onboard a radio controlled aircraft. The voltage onboard will start at 12.6v and run down to about 9.9v and I'd like the camera to always see 12v. So I think I'm limited to a buck boost type switching regulator and am just concerned about the noise... what exactly that means and how I may address it.

    Thanks.
     
  2. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    Normally additional LC filter gets rid of most of this noise.
    It is also possible to use extra linear regulators for some parts of the circuit.

    22uH together with a small capacitor could be used. Depends on the switcher IC.
     
  3. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    Switchers have noise due to the necessary switching of the current rapidly on and off to perform the voltage conversion. This generates a small ripple voltage centered at the switching frequency, which can be suppressed by additional filtering on the output if its a problem. They can also radiate high frequency harmonics of the switching frequency which can affect circuits sensitive to high frequency pickup.

    You can try a switching regulator for you application and, if there is a noise problem, add some additional LC filtering on the output.

    I once had to design a 5V,1A switching regulator with no more than 12μV (!) of output noise. I achieved that with an input common-mode choke, an additional LC output filter, and mounted it in a EMI sealed box with feedthrough filters on all inputs and outputs.
     
  4. ScottWang

    Moderator

    Aug 23, 2012
    4,855
    767
    I am curious what kind of the loading needs the noise so tiny?
    The standard ripple of the liner power about 0.1%, especially your power is using switching mode.
     
  5. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    It was power for a microwave PLO (Phase Locked Oscillator) which had very low spurious noise requirements and was quite sensitive to any noise on the power rail.
     
    ScottWang likes this.
  6. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Its the harmonics I'm mostly concerned about. Ripple I can deal with (filter), but I'm worried about radiated rf noise wreaking havoc with the nearby transmitter (900Mhz) and/or receivers (433Mhz, 1227.60 MHz, 1575.42Mhz, 2400Mhz). I can reduce efficiency slightly and get the switching frequency down to 50kHz... is 17 harmonics away "safe"? I'd suspect so. I don't have a spectrum analyzer and real-world testing could be... well... dissappointing :)
     
  7. crutschow

    Expert

    Mar 14, 2008
    13,016
    3,235
    If you are worried about harmonic radiation, then you may need to place the converter in a sealed metal box. If that's not sufficient then you may have to add feedthrough filters. Also a common mode choke on the input and/or output reduces conducted harmonics on the connecting lines, which can also radiate.

    It's difficult to determine how much shielding and filtering you need without testing with the transmitters and receivers that will be nearby.
     
  8. JMac3108

    Active Member

    Aug 16, 2010
    349
    66
    Radiated noise can be a problem with switchers but can be fixed with some attention. I've spent many days and weeks fixing radiated emissions problems on switchers.

    Mostly the fixes amount to slowing down the edges of your switching waveforms and snubbing switching transients. These can cost you a few points in efficiency but are necessary if you can't pass the required FCC radiated amissions testing.
     
  9. JMac3108

    Active Member

    Aug 16, 2010
    349
    66
    Do you have flexibility to change the switching frequency? Move it to a place where the harmonica don't cause a problem?
     
  10. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Yah. My initial question was just really to get a handle on what I'm looking at. Looks like I'll just have to give it a try and see what I see. Thank.
     
  11. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    There are exactly two kinds of noise created by a switcher: conducted noise and radiated noise. Conducted noise is what is passed in the lines (input/output leads) and radiated noise is what is radiated off into free air.

    NOTE: any unshielded conductor which is carrying conducted noise will radiate noise in the form of e-field.

    Magnetic components often radiate magnetic flux which is another type of radiated noise.
     
Loading...