Remove DC Bias in Precision Rectifier

Discussion in 'The Projects Forum' started by bonchenko, Apr 11, 2013.

  1. bonchenko

    Thread Starter New Member

    Apr 11, 2013
    2
    0
    Hello, my name is Bontor. I am trying to measure current using Hall effect sensor ACS758. Its output have DC offset 0.5*Vcc (I use 5Vcc). The resulting signal is AC signal with 2.5V DC bias.
    I am trying to convert the AC signal to DC using precision rectifier circuit in here (http://sound.westhost.com/appnotes/an001-f8.gif). The problem is, with DC bias, the output will not rectified. If I am adding capacitor in series with the input, the bias is still exist (only reduced)

    Do any of you have suggestion as how can I get DC output (with voltage = Vpeak of the AC signal) from the Hall effect sensor? I am going to read the value using Xbee, which is a radio module that cannot be programmed. Therefore, I am planning to read just the DC value and calculate RMS by dividing it with 1.4 (square root of 2)
     
  2. #12

    Expert

    Nov 30, 2010
    16,310
    6,817
    The first answer is: Use a power supply that has both a positive supply and a negative supply, and rectify from 0V DC input. Because nobody wants to do that, rail-to-rail operational amplifiers were invented so you can input to a 0 volt level at your precision rectifier.
     
  3. bonchenko

    Thread Starter New Member

    Apr 11, 2013
    2
    0
    I have already done that, if I generate 22 mVpeak signal using signal generator, 50 Hz without DC bias, the output is rectified (tried on two circuit, one using op amp with positive and negative supply, second circuit use single supply). It works.

    The real problem is that the sensor (ACS 758) automatically add DC bias in its output. How can you remove the DC bias without cutting down the output voltage?
     
  4. #12

    Expert

    Nov 30, 2010
    16,310
    6,817
    I did not read carefully enough. I will post a circuit that can be set for any DC offset in the positive and have independent AC gain. I have not customized it for you. Is it enough for you to finish the design for your needs?

    and..another idea where the input DC creates it's own compensation voltage. This amplifier chip has a rather large input bias current compared to modern chips. The filter on the negative input could be a lot higher impedance if you change the amplifier number.
     
    Last edited: Apr 11, 2013
  5. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Do you really want to see the FW rectified sine wave, or do you want the output to be peak rectified DC?
     
  6. #12

    Expert

    Nov 30, 2010
    16,310
    6,817
    It doesn't matter. X number of peaks per second or 2X number of peaks per second will read the same on a peak detector.
     
  7. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I know that, but how is that relevant to my question? The implementation for a peak detector is different than for a rectified but unfiltered output. I thought that if he wants DC out, we could maybe design something for that. Of course, you can put a LPF on the rectified output, which might be OK.
     
  8. #12

    Expert

    Nov 30, 2010
    16,310
    6,817
    The OP covered your question in post #1.
    I am not trying to be relevant to your question. I am trying to be relevant to the OP's question.
     
  9. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I don't mean to nitpick or be argumentative. I was just trying to resolve the difference between his statement:
    and his link to a schematic.
    I still am not sure which he wants, but I would guess it is his statement, and not the circuit in the link.
     
  10. charlielaub

    New Member

    May 8, 2013
    1
    0
    I have found that the circuit shown in Figure 6 on Rod Elliot's page about precision rectifiers to have very low DC offset and it uses very low cost op amps:
    http://sound.westhost.com/appnotes/an001.htm

    If you add a capacitor to the circuit before R1 it will block the DC portion of the input signal. 1u will give you a 15 Hz cutoff frequency, 10u 1.5 Hz, etc.

    -Charlie
     
Loading...