Converting 0-200millivolt ac to 0-5volt dc

Discussion in 'Analog & Mixed-Signal Design' started by just_learning, Oct 5, 2016.

  1. just_learning

    Thread Starter New Member

    Oct 5, 2016
    hey all,

    I am currently trying to make use of a signal produced from a vibration sensor ( unfortunately not much info on the internet to support it ).
    The vibration sensor is a accelerometer type and gives a voltage of 0-200millivolts ac out. I am planning to convert this voltage to a more useful 0-5volts out to be able to convert it to a standard 0-20milliamp analogue signal. I am not sure if an instrumentation amp is the answer or if an LM op amp would work.

    Any assistance will be greatly appreciated
  2. Sensacell

    Senior Member

    Jun 19, 2012
  3. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    The GBW or Gain Bandwidth Product matters. IA's are typically used for lower frequency signals. The frequency of interest needs to be known here too. Some sensors are centered around Vcc/2 and I might expect accelerometers to have an AC component. Generally, you can't get exactly zero and exactly Vcc from a single supply OP amp.
  4. AnalogKid

    AAC Fanatic!

    Aug 1, 2013
    You need three separate circuit functions, although some of them can be combined. In the usual order of things, uou need gain (is the 200 mV peak, peak-to-peak, or RMS?), you need an active rectifier to convert the AC signal to unipolar pulsating DC, and you need a filter to smooth that to a DC level than can modulate the 0-20 mA driver.

    What power sources are available for the circuit?
    Is the accelerometer single-ended or a bridge type?

  5. ci139


    Jul 11, 2016
    i assume it's a random vibration - not the fixed frequency with varying amplitude
    in latter case you could tell from rising edge slope and phase the amplitude (is apx. ̶F̶M̶ ̶ AM-ing f.fix with your input)
    in the first case the filter has inertia
    - just pointing out - not suggesting anything

    actually i analyzed what i first thought was the energy transfer to monotonic
    and there seems to be 2 different solutions
    1. to pass the input through spectrum analyzer - sum outputs
    2. if the vibration is single wave then there - perhaps - is a way to predict the amplitude by progressive differential error reduction as A1f1=A2f2 e.g. the differential progression for different amplitudes×frequencies won't correlate
    Last edited: Oct 5, 2016