time constant of integrator

Discussion in 'The Projects Forum' started by albertino, Nov 3, 2009.

  1. albertino

    Thread Starter New Member

    Apr 9, 2008
    Good afternoon to everyone,
    I have got a quick question for you guys.
    I'm going to build a delta modulator for my final year project.
    The signal into the modulator will be a sine wave of 3khz (to simulate human voice) and after passing through a comparator and a d type flip flop, it will be reconstructed by an integrator (basically it's an active low pass filter with a resistor into the negative input, a capacitor and a resistor in parallel on the negative feedback loop and two identical resistors to form a voltage divider rule connected to the positive input ( i'm looking to another circuit and considering that the op amp will operate with one supply, i think this is done because I will be working half way between +Vcc and ground).

    I have seen on some projects on the internet that a time constant of 1ms will be enough to reconstruct my signal.

    My questions are:

    Is the time constant the product of the resistor and capacitor in parallel on the negative feedback loop? Do i have to include even the single resistor on the negative input? If yes, why???

    Why, for a range of frequency between 300 Hz and 3Khz, a time constant of 1 ms is enough to reconstruct the signal?? What is the right way of reasoning and decide the right time constant knowing the maximum frequency of the system to be reconstructed???

    Thanks in advance


  2. russ_hensel

    AAC Fanatic!

    Jan 11, 2009
    An integrator does not really have a time constant: it has a slope, (volts per sec ) per volt. Not sure about the rest of the question. You verbal discription of the circuit is too much to wade thru: a picture ( skematic ) is wourth .001 megWords.