Post Go back to editing

Reducing noise in a simple current measurement

I need to measure the impedance across a specialized electrochemical cell with as low noise as possible, for a research project. The impedance of the cell is about 30 Mohms, and the potential across the cell is a 50 mV 500 Hz sine wave. The circuit measures the current, from which the impedance can be determined.

I am using an AD797 op amp with a carbon 1 Mohm resistor in the feedback loop. Can the noise of the output signal be reduced by using a different op amp? What about replacing the feedback resistor with a metal film resistor? Is there a better overall circuit design that would provide lower noise? I am seeing a noise level of about 100 Kohms in the 30 Mohm measured impedance, and need to reduce the noise as much as possible. 

Parents
  • Hi,

    Just want to confirm some items and you may agree.

    The objective is to measure the Impedance of Electrochemical cell, so the 30Mohms here is an average or a datasheet value?

    Given the 30Mohms and 50mV, the current will be 1.67nA which is too small as compared to the Bias current of AD797 which is around 0.9uA.

    This means that the Vo is more on Bias current dependent rather than Input Current.

    Try to check and consider ADA4530, for low Bias Current.

    For the 1Mohm feedback resistor, I just wanted to know how it was derived.

    The Feedback Resistor value is usually taken from the Max Voltage swing divided by the max Input Current.

    We have also an Application Note which may help you with your design.

    https://www.analog.com/media/en/technical-documentation/application-notes/AN-1563.pdf

    Hope this helps.

    Thanks.

  • Hi Fdelaram,

    Thank you for your response. The electrochemical cell is in a non-standard configuration, for use in research. The impedance has been empirically determined to be around 30 Mohms. We will try the ADA4530. The 1 Mohm feedback resistor was selected to attain as much gain as possible with a stable measurement.

  • Hello Chris12345,

    I assumed that the 50mV is fixed, so I treated  your circuit as TIA.

    Your input current will be now 50mV divided by the Impedance which is 30M in your case.

    Your Vo will be the input current multiplied by the feedback.

    In this example, the Vo is relatively small when feedback R is 1M.

    Again, this approach is on TIA.

    Kindly proceed with the ADA4530 and if you can share the result, it will be much appreciated.

    Thanks.

  • Hi Fedelaram,

    Using ADA4530 helps somewhat. I increased the resistance in the feedback loop to 7.5 Mohms, but any more than that and the output starts having a large offfset with reduced signal. In order to achieve additional gain, I added in another op amp, an AD797, with a 100 ohm input resistor and a 1000 ohm feedback resistor. This gives a gain of 9.1x. However, increasing that feedback resistor above 1000 ohm also causes the output to have a large offset with reduced signal.

    Any idea why increasing either of the feedback resistors would cause the output to suddenly have a large offset? This seems really odd to me. With the existing feedback resistors, the offset is unnoticeable, and the noise level is about 1600 ohms for the 30 Ohm impedance measurement. Still a little high, really need to get it down to about 100 ohms noise. I already do digital filtering of the output signal. Maybe also try thin-film resistors?

  • Hi Chris,

    As checked, you have used +/-12V supply for the ADA4530.

    This exceeds the Abs Max supply for this part. Any result is not guaranteed and may permanently damaged your amplifier.

    Please try to lessen the supply first and let's check the response.

    Also, we have LTC6268/LTC6269 which has a low input bias current as well.

    Thanks.

Reply
  • Hi Chris,

    As checked, you have used +/-12V supply for the ADA4530.

    This exceeds the Abs Max supply for this part. Any result is not guaranteed and may permanently damaged your amplifier.

    Please try to lessen the supply first and let's check the response.

    Also, we have LTC6268/LTC6269 which has a low input bias current as well.

    Thanks.

Children
No Data