0
\$\begingroup\$

Why is it not popular using common mode chokes in differential mode configuration for differential mode noise filtering?

Mostly when using fly-wire type chokes with filters, where the chokes are simply kept inside the enclosure with its wires connected to X and Y capacitors placed on the PCB (due to its size which makes it costly to be placed on board), can't this kind of arrangement help avoid placing one common mode choke and two separate single winding chokes and go with one common mode choke for common mode filtering and another common mode choke connected in differential mode configuration as shown in attached schematicsenter image description here?

Could saturation be a problem for low current design? If yes, I dont understand why not if two separate chokes are used. I am assuming the core is properly sized to not go to saturation during normal operation.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Try simulating the scenario you have drawn above and you will be in a better position to see why things are done the way they are done. \$\endgroup\$
    – Andy aka
    Nov 13, 2019 at 12:19
  • \$\begingroup\$ I've seen this done in higher power applications. The permeability of the core really matters when doing this. But as @Andyaka recommends, simulate it to see what the trade offs are. \$\endgroup\$
    – Aaron
    Nov 13, 2019 at 17:28

2 Answers 2

1
\$\begingroup\$

Differential signals are often used where EMC is a concern, as interference is usually common mode and a differential signal on a close pair will radiate less. The EMC enemy is more often common mode noise or emissions. Your suggestion would be more useful in situations where you need multiple signal paths and want to do as much as possible to ensure they're all identical.

\$\endgroup\$
2
  • \$\begingroup\$ Hi Cristobol, I believe you are talking about EMI induced on small signal lines. My primary goal is to mitigate lightning induced EMI affecting our SMPS power supply which, in my experience has high energy pulses of both common mode and differential mode nature. Do you find any reason why this wouldnt work? I will run a simulation and publish my results here.Thank! \$\endgroup\$ Nov 14, 2019 at 3:35
  • \$\begingroup\$ @JaacsPottas LOL, I don't deal much with lightning, class 4 ESD is more my speed. I'll have to take your word on the requirements. \$\endgroup\$ Nov 14, 2019 at 13:37
1
\$\begingroup\$

The common PI filter uses shunt caps to ground for CM suppression and a differential Cap for differential noise suppression. The % of suppression depends on the match tolerances of L1a/L1b and C1a/C1b and the impedance ratios, for an imbalance CM noise creates a differential noise from the difference.

Magnetics rarely operate well (linear) over 2 to 3 freq. decades due to both self-winding capacitances at the high-end or insufficient impedance or mutual coupling at the low end. So multiple chokes may be used with different inductance and cascading impedance increases for EMI reduction. Yet Chokes may be useable over 3.5 f decades and cores are used with low hysteresis.

Since passive LC low pass filters often rely on Specs for DC current or 50 Ohm T lines, DC lines or ACs they tend to be lower inductance and primary rating is current.

Common mode chokes may crossover in the range of DM chokes which are designed for tight balancing and assists the CMRR of a CM impedance from a CM source noise to a pair of CM shunt attenuation capacitors. The tolerance error of the C pair also affects the CMRR of the LC filter pair and the load R dampens the resonant gain.

So in short, CM chokes tend to be better matched and have higher inductance for each leg to achieve a minimum impedance at some test frequency range. While DM chokes are lower impedance to match the load and thus be well dampened at the breakpoint frequency.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.