Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Phase detector for PLL with low SNR

  1. Nov 17, 2013 #1
    I'm making a system with PLL. Here's more or less a principal scheme. What happens here is that the signal in this compensating system becomes really low and phase detector might not work properly. What I want to ask, what type or how to make a phase detector which can detect phase correctly from a very low signal. Any scheme idea?

    Attached Files:

  2. jcsd
  3. Nov 17, 2013 #2

    Can you supply some more detail? Are the signals digital or analog? When you say "low signal" do you mean you wish to detect very small phase error, or that the amplitude of the signal is small.
    Last edited: Nov 17, 2013
  4. Nov 17, 2013 #3
    The system's purpose is to measure a surface potencial with vibrating zond of a material. The signal is analog and measured with voltmeter. PLL circuit is used to automatically adjust frequency to a resonance f. Because frequency is at resonance the phase is very unstable. By saying low signal I mean the amplitude is small and it kinda dissapears in the noise which makes it hard to detect. Any particular questions?
  5. Nov 17, 2013 #4
    Why don't you use the output of your generator to drive two synchronous detectors but maintain a phase difference between the generator inputs to the two synchronous detectors of 90 deg? This is typically referred to as I & Q demodulation.

    You would take the analog outputs of both synchronous detectors, square the values and add them together. The sum represents the amplitude of your signal. Note that this method works even if neither phase from the generator is in phase with your signal. The effect of noise can be minimized by the use of a low pass filter on the output.

    To accomplish the squaring function you can either use A/D converters to convert to digital values which you square and then either convert back to analog or use the digital values. Or you can use an analog multiplier such as the AD633ANZ to square the analog values directly.

    If you look around you may be able to find ICs that do the I & Q demodulation for you.
    Last edited: Nov 17, 2013
  6. Nov 18, 2013 #5


    User Avatar
    Science Advisor
    Gold Member

    This doesn't make sense to me. Since the slope dphi/df is maximal on resonance, doesn't your feedback loop have maximal gain there? I'd expect best stability on resonance.
  7. Nov 18, 2013 #6
    I have to say I'm a little confused as well. The phase should be fairly stable at resonance if the Q is decent. At the very least, it sounds like you're going to need an amplifier with a bandpass filter to get enough signal to drive your phase detector.
  8. Nov 19, 2013 #7
    any small frequency deviation from the resonance makes a phase to change from +90 to -90. So I need f to be stable at resonance and phase to be stable. My lecturer said that the gain will be minimal.
    Honestly, I have no idea what I'm doing here.
  9. Nov 19, 2013 #8
    Phase with respect to what - your generator?

    How is f created? If you have a pickup, what kind of pick up is it? Are you able to amplify and filter f before you try to phase lock to it?

    Can you clarify why your instructor instructor said gain will be minimal? Can you give us a better description of your setup?
  10. Nov 19, 2013 #9

    That sort of phase shift around resonance is typical. The swing thru zero is what marks the resonant frequency.
  11. Nov 19, 2013 #10
    I have to ask, why are you using a PLL for this? If all you want is to provide stimulation to a resonant device, all you need is a simple signal generator tuned to the resonant frequency.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook