Hardware filtering low kHz noise wave from cable?

Discussion in 'General Electronics Chat' started by antseezee, Jul 3, 2008.

  1. antseezee

    Thread Starter Active Member

    Sep 16, 2006
    45
    0
    Hey everyone. I'm working on a project that includes a sensor clip with a photodiode and LED. I flash the LED at a rate of 1kHz with a pulse wave fed through the sensor wire. This 1kHz pulse wave is appearing as a 1kHz noise wave riding on other equipment measurements.

    Is there anyway I can hardware filter (shielding, coating) out this 1kHz noise and prevent it from propogating to other equipment measurements?

    The only solution I have come up with was a notch filter and it works. But I'm looking for a cheapier solution. I've tried ferrite cores to place on the cable, but they're designed for such high frequencies. Any help appreciated - thank you.

    P.S. I tried to EMI shield the cable, but since it's a sensor clip and not in an enclosure, this was impossible.
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Hello,

    How is your powersupply decoupled?
    It is the best way to have an 100 nF capicator on each IC from ground to the +Vcc and -Vcc as close as possible to the IC.
    A signal can travel along the powelines when this is not the case.

    Some more information can be found over the links from this page :
    http://www.educypedia.be/electronics/noise.htm

    More topics on this site are :
    Miscellaneous topics: Barcode Chip packaging technology ESD Miscellaneous topics Semiconductor Manufacturing Decibels Grounding and shielding Noise Soldering EMC - EMI Magnetic stripe technology PCB design Transient suppression

    Greetings,
    Bertus
     
  3. antseezee

    Thread Starter Active Member

    Sep 16, 2006
    45
    0
    My power supply is decoupled with a 100uF capacitor for +5V and 100uF capacitor for -5V in parallel to ground right where the output buses begins. I created a dual source power supply (+/- 5V). Every major IC on my board (except charge pumps) has a 0.1uF cap parallel to ground. There is an individual cap for both the +5V and -5V supply on each chip.

    To make matters worse, I confirmed the noise propogating from the sensor clip. With the sensor clip off (but the circuit on), the 1kHz noise wave is not visible. When the sensor clip is turned on, the 1kHz noise wave suddenly appears. The medium the object is placed on is a small animal where the other instruments take measurements.

    Like I said, I'm not too familiar with the electro-mechanical aspects of filtering stuff. I'm trying out some EMI shielding tape (which is suppose to prevent EMI interference, you wrap it around cable joints). The noise only appears where the LED & photodiode make contact with a medium.

    I'm wondering if there's some stray capacitance effect coming from the photodiode, or some LED noise caused by my pulse wave. Like I said, I built a portable notch filter that eliminates the noise with BNC connectors, but there must be a cheaper hardware method of fixing this solution?

    That link was helpful - thank you.
     
  4. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Hello,

    Is there a 1 khz oscillator in yous circuit? (a block has many harmonics)
    If yes? Is this a blok voltage oscilator or sinewave?
    What is the amplitude of the oscillator?

    Greetings,
    Bertus
     
  5. antseezee

    Thread Starter Active Member

    Sep 16, 2006
    45
    0
    There is a 1kHz oscillator output being produced by a PIC. It is a square-wave, 0 to 5V signal with 30% duty cycle. I was hoping there would be a way to isolate this problem by placing some sort of shielding over the photodiode and LED to prevent it from spreading to other instruments.
     
  6. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,347
    Last edited: Jul 5, 2008
  7. antseezee

    Thread Starter Active Member

    Sep 16, 2006
    45
    0
    Thank you for your help Bertus. If anyone is interested, I found a partial solution to this problem. Obviously, you cannot place a sensor clip inside an enclosure if you are trying to attach a small animal for measurements. To eliminate noise from propogating to nearby sensors from the clip itself, I found EMI shielding tape. It's basically copped-coated tape that you angle around cable joints or shape to isolate noise in portions of circuits.

    I sampled a roll of this product and it attenuated the noise signal by approximately 60-70%. Great stuff, and not expensive by any means.
     
Loading...