0
\$\begingroup\$

I have a switched mode dual power supply with +12V, GND and -12V outputs. I'm seeing a lot of high frequency noise on the output of this power supply that I'd like to attempt to filter out. The noise appears as somewhat random bursts of 12MHz oscillations and these bursts appear as frequently as around 12.5us intervals (about 75kHz). The bursts are about 1Vpp in amplitude.

I've attempted to use RC filtering and LC filtering. The RC filtering did not achieve much (although I did use an electrolytic cap which may have high ESR), but I was able to cut the amplitude of the noise in half using a 33uH inductor and a 47uF capacitor. Using a common mode choke, I was able to reduce the noise even further, to about 0.15Vpp. This was all measured when drawing a constant 0.1A through the +12V output of the SMPS to ground.

So the common mode choke seems pretty effective at reducing noise from the SMPS. However, I'm not sure if I would be able to use one effectively with a dual power supply. Would I use two common mode chokes, one between +12V and ground, and another between -12V and ground? I don't believe I can guarantee the current will be balanced on both sides of the CMC. Some current paths will go from +12V to gnd, some will go from gnd to -12V, and some will go from +12 to -12V. Can I even use a CMC effectively in this scenario?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Sure, the principle works with any number of windings. The selection of parts is not as great though for 3 or more lines. It won't guarantee balance on the +/- 12V lines (vs GND) though, it would only block (or present a fairly high impedance to) the net current between all 3 lines \$\endgroup\$
    – Pete W
    Commented Apr 3 at 0:37
  • \$\begingroup\$ Also sounds like potential confusion whether you're really trying to reduce the differential mode noise of the SMPS? If a CMC improved the 12V vs GND, there's something else going on too, maybe to do with input, output, and measurement ground paths \$\endgroup\$
    – Pete W
    Commented Apr 3 at 1:50
  • \$\begingroup\$ @PeteW Thanks for the info about CMCs with multiple windings. I haven't heard about these so I'll look into them more. Would you have any advice for investigating what else might be going on here? My measurements seem to be OK as I am able to take measurements of a +5V SMPS given the same inputs (+24V and ground) as the +/-12V SMPS and it presents very little noise in comparison. All my ground paths are connected. \$\endgroup\$
    – Jordan Melo
    Commented Apr 3 at 2:19
  • 1
    \$\begingroup\$ if you put up a schematic you'll prob. get a lot more help \$\endgroup\$
    – Pete W
    Commented Apr 3 at 6:15

0

Reset to default