0
\$\begingroup\$

In order to properly deal with EMI interference problems one needs to understand where/how the offending currents flow. With grounded balanced audio equipment, for example, common practice is to shunt RF common-mode noise that makes its way into the signal through cables to the chassis, where I always assumed it traveled to earth through the ground cable. I viewed an earth connection as a giant sink for this current, thinking that for whatever reason, RF common-mode current picked up by cable shields etc tries to find the fastest path to earth ground, and doesn't return to the source like it would with normal differential mode currents. But aircraft and ground vehicles don't have an earth ground connection, so clearly that isn't correct. So how and where do these RF CM currents flow when they are shunted to the equipment or vehicle chassis? Do they have the tendency to want to return to their source and flow in loops or just dissipate?

Update: Anybody??

\$\endgroup\$
7
  • 1
    \$\begingroup\$ "aircraft and ground vehicles don't have an earth ground connection" sure they do, it's capacitive, but it is still there. \$\endgroup\$
    – Trevor_G
    Nov 24 '17 at 19:19
  • \$\begingroup\$ Ok, I could see the ground vehicles thing, but aircraft 5 miles up? Are you saying my assumption of RF currents wanting to sink to earth ground is correct? Because I was told in another post earth ground has nothing to do with RF common-mode current flow \$\endgroup\$
    – User7251
    Nov 24 '17 at 19:29
  • 1
    \$\begingroup\$ Ultimately the return currents for the RF signal you are picking up has to make it back to wherever it came from. Whether that be through your chassis, or via capacitive coupling to the ground, or clouds, or whatever, it has to get there. Of course if the source of the RF interference is inside the aircraft, that's a different story. \$\endgroup\$
    – Trevor_G
    Nov 24 '17 at 19:33
  • \$\begingroup\$ I think some of my misconception about the earth return comes from thinking of induced CM voltage from switching supplies and RF common-mode interference as behaving the same. So you're saying that the CM current flows in complete circuits back to the source just like differential mode currents? I'm envisioning transmitted radio waves being picked up as common-mode on aircraft wiring and being shunted to the aircraft chassis, you're saying it then radiates back to the source transmitter? \$\endgroup\$
    – User7251
    Nov 24 '17 at 20:08
  • \$\begingroup\$ An in-depth answer with some examples of how this works would clear up the confusion. \$\endgroup\$
    – User7251
    Nov 24 '17 at 20:10
1
\$\begingroup\$

Everything is differential mode if you look at a big enough picture — there are no circuits mysteriously lacking a return path here. The incoming radio wave causes a potential — a voltage difference — between two separated conductors. Which two conductors? Whatever ones are available!

When you are concerned about shunting RFI to earth ground, it is because there is a potential difference between earth and that part of your circuit. But this is not a special property of earth — it's just playing the role of one of the conductors. In an aircraft, the potential difference must be between two different parts of the aircraft. Any two conductors — or parts of the same conductor — having some significant physical separation can function as a receiving antenna in this way.

The goal of your shunt is not to send the interference specifically to earth in a somehow one-wire fashion; it is to give the interference arriving on some two conductors a short circuit between them — one shorter than the path it could instead take through your sensitive electronics if the filter were absent.

\$\endgroup\$
2
  • \$\begingroup\$ This is a really great answer, thanks. So when designing for EMI rejection, is it best to assume every cable coming into the device is essentially the + pole of a voltage source, and any other cable is a - pole, and you want to create as short a path as possible (diverted around the signal, through the chassis) between these points? Is it best then to just create a shunt to chassis at every entrance point (power/IO/etc.) of a susceptible device with a metal enclosure? \$\endgroup\$
    – User7251
    Dec 1 '17 at 0:01
  • \$\begingroup\$ @User7251 My own knowledge is not enough to confidently agree with such broad statements. \$\endgroup\$
    – Kevin Reid
    Dec 1 '17 at 2:03

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.