PCB equivalent of a wire through a ferrite ring?

Discussion in 'General Electronics Chat' started by s_mack, Sep 14, 2012.

  1. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Hi. I'm designing a device which interfaces with an existing device. I take two toggle switches on the existing device to my board and process them as simple inputs on an AVR. They are debounced. The problem is that these switches are huge, cheap, clunky examples and some (not all) create a lot of EMI. It is handled just fine by running the suspect wires through a ferrite loop.

    But I don't want to run the wires through a ferrite loop. And I don't want end users who don't know if they'll need one or not. I'd rather solve it on-board if I can.

    So can I run the traces that those wires connect to through a PCB version of a ferrite loop? i'm not sure what that would even look like :)

    I've tried a ton of google searches but I keep getting info relating to switching power supplies so I'm obviously not sure what to search for.

    I'm hoping for a very inexpensive solution to the problem. Ideas?
     
  2. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
  3. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    I have, yes. I haven't found anything related to its use for signal noise suppression though. Its always power bypassing, etc. Typically dealing with noise generated by switching regulators.

    Is just placing a bead inline with the signal similar to wrapping the wire around a ring?
     
  4. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Switches don't create much EMI unless they are switching a substantial amount of current. If they are on long wires, they can either pick up EMI, or radiate it if it is present on your board.
    Board-mounted beads can be as effective as discrete toroids, but not all beads (or toroids) are created equal. You need to select the bead to match the application.
    Having said all that, I have to say I don't have a lot of experience with either part, but I generally understand the basic theory behind them.
    This looks like a good treatise on the subject.
     
  5. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Yeah, I read that already... again, its about power.

    I dunno why its creating so much EMI then. They aren't switching any appreciable amount of current. Its simply taking an input to ground momentarily. I swore it was software related, but on extensive testing, rapid flipping of the switch will ALWAYS lock up the CPU without the torroid and will NEVER lock it up with the torroid. 100% repeatable over thousands of tests.

    While curiosity sure has my interest piqued... practicality states that we're better off dealing with the symptom here rather than trying to chase down the cause.

    Anyway, I'll just get a bunch of general purpose EMI suppression stuffs (beads, pass-through caps, specialized ICs, etc) and see what works I guess.
     
  6. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Do you have an Oscilloscope?
    It's a lot easier seeing what's happening,rather than guessing.

    How momentarily? If it is for a reasonable time,you could wire a small capacitor across the switch to slow the rise time & get rid of any high frequency transients.

    Are your wires wound as a pair on the toroid,in the manner normally used to suppress common mode signals,or is just the "hot" lead wound on the toroid?

    If it is the latter, the toroid is just acting as a series inductor,& can probably be replaced with a discrete series inductor ,shunt capacitor ,or both.
    Maybe another option might be a series resistor & shunt capacitor.
     
  7. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
    Maybe the inductance of the long wires is creating a momentary high voltage while switching. Is it possible that a couple of schottky diodes to the power rails would solve it?
     
  8. atferrari

    AAC Fanatic!

    Jan 6, 2004
    2,648
    764
    Could you show how, Mark?
     
  9. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
  10. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    more on the wires:

    They aren't really "long", but that's a relative term, isn't it :) Its about 5" and is normally twisted with 5 other wires. We unravel it for our purposes and I have no doubt that induces the problem. It is solved by wrapping that single hot wire through the torroid. I've tried also just re-twisting it back where it was "as best as I can" and it REDUCES the problem, but does not eliminate it (as the torroid does). Its also not always a problem in every case. It seems to depend on the end-unit's particular switch and wiring. But the ones that do exhibit the problem, exhibit it consistently.

    I have already got an r/c on the wire at the connector and have tried many different values for the resistor up to the point of blocking the signal from effectively getting through and the EMI is still a problem at that point.

    Strange...

    I will try some of the suggestions here, thank you.
     
  11. crutschow

    Expert

    Mar 14, 2008
    13,046
    3,244
    You might try the RC directly at the switch or a small cap directly across the switch contacts (no more than about 0.01uF).
     
  12. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    Thanks for that link Mark - that looks particularly promising.
     
  13. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    You have answered my question.
    The single wire wound through the toroid creates a series inductor,which is a high impedance to high frequency transients.
    Try a discrete inductor,or several.
     
  14. s_mack

    Thread Starter Member

    Dec 17, 2011
    187
    5
    crutschow - I can't do anything at the switches. My board is interfacing with others' board. Other than unplugging their wire and putting it into ours, no modification on their end is practical.
     
  15. crutschow

    Expert

    Mar 14, 2008
    13,046
    3,244
    OK, that leaves that out.

    You might also try using a shielded (coax type) wire for the offending wire, if that's possible, since moving the wire seems to affect the transient. Be sure the shield is connected to a good ground with a short connection to the circuit.
     
  16. JMac3108

    Active Member

    Aug 16, 2010
    349
    66
    S_mack,

    Not sure where you're getting your info on surface mount ferrite beads. These are used in industry extensively for EMI suppression of I/O signals (and power of course). Check out Murata BLM series. They're available in various sizes and you can get them at DigiKey.
     
Loading...