Connect with us

RF Interference caused by spark gap!

Discussion in 'Electronic Design' started by Steve Minshull, Dec 9, 2003.

Scroll to continue with content
  1. I'm trying to count sparks from a spark gap using a simple PIC uC circuit.
    I have a small loop of wire picking up the RF interference with a rectifier
    and 5V clamp across the loop. The signal then goes to a uC input. All the
    circuitry is inside an earthed diecast box with only the wire loop
    protruding out of it. When I tested the circuit the uC would not do
    anything sensible (Probably being reset on each RF pulse) The box seemed to
    radiate the RF inerference. A scope probe picked up the switching waveform
    when it was near the box and everywhere in the circuit.

    Has anyone got any advice for calming this RF interference down so that the
    system will function?

    Cheers in advance

    Steve Minshull

  2. It has been my unfortunate experience as well that PIC microcontrollers are
    tremendously vulnerable to RFI/EMI upset. I hope they address this weakness
    in their future products (especially their microcontrollers or DSPs
    specifically targeting power electronic control).

    In the meantime you can keep working on trying to shield your
    microcontroller circuit better (also perhaps trying to shield the spark gap
    as well). Is the high voltage totally galvanically isolated from the PIC
    circuitry? If it is, then I might suggest trying to insert some
    intermediate circuitry including an optocoupler between the loop antenna and
    the PIC. Good luck, I'm sure this is a most frustrating problem.
  3. default

    default Guest

    You don't mention how it is powered. The optimum solution there is to
    enclose the battery with the PIC

    Some signal massaging might be called for. Use a small transistor amp
    and limit the bandwidth with some caps.

    If all else fails, feedthrough caps or filters work wonders.

    Scope readings lie. The scope body itself is a large antenna, the
    "Earth" connection (like the earth on your box) is just a wire with a
    lot of inductance. The inductance isolates the high frequency
    component you are trying to measure.

    Your scope probe may have a ground ring on the tip, if possible use
    that in preference to the longer 3-4" lead when grounding to the
    circuit. Use the scope to ground your box while taking measurements -
    you can't have two grounds without a ground loop - the ground loop is
    another loop antenna (Duh!)

    Other than trying the pic in a box, what have you done? This isn't
    the fault of the processor - this is an application problem.

    I have a lightening detector (detects sparks from a piezo lighter for
    a test of the device). To keep the circuit from counting multiple
    strikes (a spark is not a single event, but a series of decaying RF
    waves) I amplified the signal with a single transistor broadly tuned
    to 100 KHZ with an inductor capacitor tank, then use that to trigger a
    555 monostable. Once triggerred, the monostable ignores the rest of
    the wave and outputs a clean pulse into my counter.
  4. How far away is the spark? I'm trying to get an idea of the emission
    energy. If you have a spectrum analyzer (heck, even an AM / SW radio)
    Try to get an idea of where the spark energy is concentrated
    spectrally. Tune your pickup loop with a capacitor to that frequency.
    The energy from the arc may be very broadband, just look for where it
    has a lot of energy relative to the background EM.

    Are the sparks supposed to occur within a particular interval, or is
    it more random, such as the Partial Discharge within a power
    transformer? (I've always liked that euphamism).

    The rectifier should be a low-capacitance type, e.g. PIN diode, or RF
    detector diode, depending on the resonant frequency you choose to
    detect. (every loop has a resonance).

    Does the micro run properly with the loop winding entering the box but
    not attached to the micro? If not, the EMI may be too much for the
    micro and related layout, as you suggest. Isolate the loop
    differentially (RF transformer that works at the resonance of the
    loop) and put the detector on the secondary. This makes a balanced
    feed into the box that will be less likely to disrupt the electronics.

    Frank Raffaeli
  5. I read in that Steve Minshull <[email protected]> wrote (in <AHiBb.90$
    et>) about 'RF Interference caused by spark gap!', on Tue, 9 Dec 2003:
    Is a loop the right sort of antenna? The spark produces mostly electric
    field in the near-field region, doesn't it? A loop picks up magnetic
    field energy.
    The scope probe is an *electric* antenna, and as you have found, works
    well for picking up emissions from sparks.
  6. R.Legg

    R.Legg Guest

    You should be clamping to the supply rails, and providing a loop
    terminating impedance to get a well-defined zero.

    Place a resistor between the clamp and the IC input. 3K3 is a nice
    This limits current into the IC if and when the clamp voltage exceeds
    either rail.

  7. PIC inputs have internal diode clamps.

    However, they have a max current of about 20mA, so a series resistor sounds
    like a good idea.

    One of the microchip technotes describes a method of sensing powerline
    crossing of connecting the PIC input directly to the powerline through a
    4.7M resistor.

    Bob Monsen
  8. It is a good idea to keep those diode currents very low if you expect
    the chip to work properly. They are intended to protect the chip from
    damage, only. I have seen some very squirrelly things happen while
    passing sub milliamp currents through the protection diodes,
    especially in the A/D converter and mux section.
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day